A Corpus-Assisted Comparative Study of Chinese and Western CEO Statements in Annual Reports: Discourse-Historical Approach

2.1 Corporate Identity

Since the concept of CI surfaced in the 1950s (Lippincott and Margulies 1957), discussions and deductions have been excitingly dynamic. Until now, there is not a universal definition of CI because of its multidisciplinary nature and different emphases on the determinants of the concept. CI refers to an organization’s central, enduring, and distinctive character reflected by strategically planned and operationally applied internal and external self-presentation and behavior of a company (Albert and Whetten 1985; Birkigt and Stadler 1986). CI is defined as what an organization is (Dowling 1986), the explicit management of all the ways in which the organization presents itself (Olins 1995), an organization’s unique and interdependent characteristics and its sameness over time (continuity), distinctiveness and centrality (Larcon and Reitter 1979; Markwick and Fill 1997; van Riel 1997; van Riel and Balmer 1997). CI is a planned and operational self-presentation of a company, both internal and external, based on an agreed company philosophy (Cornelissen and Harris 1999). Moingeon and Ramanantsoa (1995) emphasizes the pattern or combination which constitutes the identity of the organization. Melewar and Jenkins (2002: 86) develops a holistic CI model and categorizes four dimensions of CI, namely, communication/visual identity, behavior, corporate culture and market conditions. Evangelisti-Allori and Garzone (2010: 12) regards CI as “a multi-faceted conceptualization of an organization’s structure, roles and values”. Okay (2013: 25–26) and Steyn (2004: 168) think CI consists of four elements, namely, the behaviors of the individuals and the organization, its communication styles, philosophy and visual design elements. Siano et al. (2013) thinks CI involves a wide range of organizational functions including corporate design and behavior, corporate culture, business strategy and philosophy of key executives. Jain et al. (2016) identifies that CI deals with strategic, emotional and social dimensions. Balmer (2017) introduces a new integrated strategic framework to the concept of CI, namely, “the CI, total corporate communications, stakeholders’ attributed identities, identifications, and behaviors continuum”. In short, the concept of CI is multi-dimensional and interdisciplinary and has received immense interest from academics of various disciplines and practitioners. Our study defines CI as the self attributes discursively presented in corporate communication (Cheng and Shi 2022: 315).

In today’s global economy, CI management process and its multidisciplinary aspects were emphasized (Bick, Marciene, and Abratt 2003: 850) and its effects on the employees’ responses were analyzed (Bravo, Matute, and Pina 2017). Nguyen et al. (2018) investigated the CI construct in the high technology industry in China. Melewar and Skinner (2018) defined and delimited the multi-modal and multi-sensory scope of CI construct and explored its applicability to SMEs. Signori and Flint (2020) introduced a linguistic methodological tool, the semiotic square to analyze the CI congruence. Işık (2021) studied university websites with content analysis methods to identify CI building of universities in the context of organizational communication. Tomo et al. (2022) studied how family firms integrated CI-building processes and socioemotional wealth dimensions when going international. Furthermore, recent studies have paid consideration to how CI is discursively constructed. Feng (2017) explored discursive constructions of CI by Chinese banks on Sina Weibo. Li (2022) explored how CI is discursively constructed in corporate communication like annual reports and corporate social responsibility reports. Comparative studies have been carried out on how CI is discursively constructed by Chinese and American energy companies (Liu and Wu 2015), Chinese and Western universities (Chen 2016, 2017), Chinese and American banks (Cheng and Shi 2022; Fu and Zhu 2022) and on the discursive strategies adopted by Chinese and American corporations (Wu and Zhang 2019). Our paper plans to carry out a comparative study of CEO statements in annual reports from Chinese listed pharmaceutical companies and Western listed pharmaceutical companies to identify the universals and distinctiveness of Chinese and Western corporate identities discursively constructed through corporate communication.

2.2 Previous Studies on CEO Statements

Previous studies have revealed the significance of CEO statements in annual reports to the putative readers, namely, shareholders or investors or society at large and have illustrated the utility as a strategic communication to enhance CI. Many studies have investigated CEO statements or letters in terms of the readability (Bayerlein and Davidson 2011; Julianaldri and Tarjo 2022; Wang, Hsieh, and Sarkis 2018), the diction and tone (Alshorman and Shanahan 2022; Blackford 2014; Craig and Amernic 2011, 2018), the informativeness, rhetoric and effectiveness as a means of strategic communication (Gâţă and Stoica 2014; Patelli and Pedrini 2014; Seo, Munehiko, and Li 2021; Yan, Aerts, and Thewissen 2019), the variation in use of meta-discourse (Huang and Rose 2018), the effectiveness along dimensions of credibility, efficacy, commitment, and responsibility (Segars and Kohut 2001), the sustainability embeddedness in the era of globalization (Arvidsson 2023) and sentiment analytics and impression management (Boudt and Thewissen 2019). Comparative studies are also an important area that has drawn immense attention. Conaway and Wardrope (2010) compared CEO statements in annual reports between United States and Latin American companies to identify similar themes. Crawford Camiciottoli (2013) explored three world-renowned Italian fashion brands: Valentino, Dolce & Gabbana and Giorgio Armani and found how companies utilized web-based communications to shape brand identity in an increasingly globalized marketplace. Khan and Sulaiman (2021) investigated the linkage between CEOs statements from five Malaysian companies’ annual reports and corporate social responsibility (CSR) reporting and uncovered the main discourses and the performativity of such communication. Liu, Bilal, and Komal (2022) compared the CEO statements in annual reports with the ones in CSR reports and found different priorities in these two related genres, with the former highlighting the economic and pragmatic concerns of shareholders and the latter the ethical ones. Wang, Liu, and Moratto (2023) compared the syntactic complexity of translated chairman’s statements (Chinese to English) and non-translated ones (English) and found translated chairman’s statements are significantly simpler in subordination as well as overall sentence complexity, providing some evidence for the simplification hypothesis of translation universals. Our paper aims to employ the Discourse-Historical Approach of Critical Discourse Studies to conduct a comparative study of Chinese and Western CEO statements in annual reports, with three distinguished Chinese listed pharmaceutical companies as one party and three renowned Western listed pharmaceutical companies as the other.

3.1 Hofstede’s Cultural Dimensions

Geert Hofstede is one of the most cited social scientists in Europe (Sent and Kroese 2022) and his empirical research framework of six cultural dimensions will be employed to explore similarities and differences in cultural values displayed in CEO statements in our study. By doing quantitative and analytical studies on work-related values through the huge data collected from more than 40 countries and regions, Hofstede identified and published the first four cultural dimensions in 1980. The dimensions are individualism versus collectivism (IDV), power distance (PD), uncertainty avoidance (UAI), and masculinity versus femininity (MAS) (Hofstede 1980). Later he added a fifth dimension to his model, long-term versus short-term orientation (LTO) (Hofstede and Bond 1988) and the sixth dimension, indulgence versus restraint (IVR) (Hofstede, Hofstede, and Minkov 2010).

Hofstede’s six cultural dimensions can be briefly defined as follows. To begin with, the IDV dimension focuses on the degree a culture reinforces interpersonal relationships and individual or collective achievement. In individualist cultures, independence, individuality, and diversity are paramount while in collective cultures, common interest, cooperation, and interdependence are greatly valued. Second, the PD dimension refers to attitudes toward differences in power and authority. Cultures with high power distance accept the fact that differences in organizational status and rank are clear-cut and that employees have a great deal of respect for those in high positions. Cultures with low power distance show a democratic system in practice and superiors are more approachable and are even challenged by employees. Third, the UAI dimension measures the extent to which society tolerates uncertainty and ambiguity. Cultures exhibiting strong uncertainty avoidance try to provide stability for their members with a great amount of detail and planning. On the other hand, cultures with low uncertainty avoidance prize initiative and consider unusual situations as opportunities. Fourth, the MAS dimension shows social role division between genders. In masculine societies, social roles are clearly distinguished and achievement, authority, money, and dominance are greatly valued. In feminine cultures, social roles are not distinctly distinguished and quality of life is a high priority. Fifth, the LTO dimension focuses on value orientation towards the future or the past and present. Cultures with long-term orientation exhibit the focus on future gains while compromising short-term ones while cultures with short-term orientation the opposite. Sixth, the IVR dimension measures the degree people satisfy the natural human drive for enjoyment. Indulgence stands for a culture where people are allowed to freely satisfy their desires related to having fun and enjoying life. On the other hand, restraint stands for a culture where strict social norms regulate the gratification of needs.

3.2 Data Sources

As we planned to conduct a comparative study of Chinese and Western corporate communication, we chose three Chinese listed pharmaceutical companies and three Western listed pharmaceutical companies. To avoid possible differences in cross-industrial annual reports, we choose companies which practice in the same industry of pharmaceuticals as our research objects. As for the three Western listed pharmaceutical companies chosen, Pfizer Inc. ranked No. 1 on 2022 global Top 25 pharmaceutical companies list, AstraZeneca plc. ranked No. 9 and Commonwealth Serum Laboratories (CSL) ranked No. 23. They are also the best pharmaceutical companies in America, Britain, and Australia respectively. Although Chinese pharmaceutical companies are in a period of rapid growth, none is on the list of Top 25. Therefore, we randomly chose three young and thriving Chinese pharmaceutical companies listed in the Hong Kong Stock Exchange as English is one of the official languages in Hong Kong and the requirements for English versions of CEOS’ statements on the website of Hong Kong Stock Exchange may be closer to the Western standards of native English speakers. Two self-built corpora were compiled of the latest five-year (2017–2021) CEO’s statements in annual reports of six listed pharmaceutical companies as follows (Table 1).

The first three are Chinese pharmaceutical companies listed in the Hong Kong Stock Exchange and their annual reports from 2017 to 2021 were downloaded from the Hong Kong Stock Exchange website by entering the stock code of each listed company (https://www1.hkexnews.hk/search/titlesearch.xhtml?lang=en). The CEO statements in annual reports from the three companies constituted the corpus of Chinese companies (Corpus CC). The annual reports from 2017 to 2021 from these three companies were downloaded from their official websites. The CEO statements in annual reports from the American, the British and the Australian companies constituted the corpus of Western companies (Corpus WC). The CEOs’ statements were extracted from the downloaded PDF annual reports, pasted to word files for checking and cleaning and saved as TXT files. For the purpose of comparison, we chose to build the two corpora similar in size: a total of 18,981 words for Corpus CC and a total of 17,287 words for Corpus WC.

3.3 Synergy of Corpus Linguistics and Critical Discourse Analysis

Based on the fundamental concepts that language is a social phenomenon and that language and culture are inseparable, Corpus Linguistics (CL) studies a large body of empirical data by employing computational techniques. Displaying word frequencies, analyzing concordance lines, and comparing semantic domains could help to uncover hidden socio-cultural patterns. Critical Discourse Analysis (CDA) conducts the systematic and explicit analysis of the various structures and strategies of different levels of text and talk (Van Dijk 2008) to reveal ideologies in society. The use of CL techniques in CDA, which is also called corpus-assisted discourse studies (Partington 2004, 2006) is becoming increasingly popular. Many a study has been carried out with a methodological synergy of CL and CDA (Baker 2004, 2006; Baker et al. 2008; Gillings, Mautner, and Baker 2023; Hardt-Mautner 1995; Krishnamurthy 1996; Koller and Mautner 2004; O’Halloran and Coffin 2004; Stubbs 1994; Yu 2019). A corpus-assisted discourse analysis is employed for the study.

3.4 Corpus Analysis Tool: Wmatrix

Wmatrix (Rayson 2008) is an online corpus analysis and comparison tool (https://ucrel-wmatrix5.lancaster.ac.uk/wmatrix5.html). Using the tag wizard in Wmatrix, text files can be annotated at two levels: part-of-speech and semantic field annotation. Part-of-speech tags are added by the CLAWS tagger and semantic field tags are added by the USAS tagger. USAS is a multitier semantic tagging system that groups words which are similar in meaning into a single category and assigns a semantic domain tag to each word in a given text. USAS classifies words into 21 major semantic categories (A–Z) (see Table 2) and 232 specific semantic fields (see http://ucrel.lancs.ac.uk/usas/usas_guide.pdf for the specific introduction). CLAWS and USAS have a high degree of accuracy applying tags to English texts, with an accuracy rate of 96–97 % and 92 % respectively. In this study, BNC Sampler CG (Spoken) Business, with 141,143 words from BNC 1994 Sampler Context Governed Business corpus (Company and trades union talks or interviews; business meetings; sales demonstrations etc.), was chosen as the standard reference corpora for key analysis and was to provide standard frequency information to compare our corpora at the word, POS and semantic levels.

With the corpus tool Wmatrix5, we identify the word frequency list, high frequency verbs, high frequency adjectives and key semantic domains in CEO statements in annual reports. Each of the observation corpora is compared with the reference corpus of BNC Sampler CG (Spoken) Business using log-likelihood (LL), which is a statistical test widely employed in corpus linguistics to measure if there is significant difference in item frequencies between two corpora (Rayson 2003). Our paper aims to identify common and distinctive categories in the observation corpora (Log Likelihood LL ≥ 6.63, p ≤ 0.01), which will facilitate the analysis of thematic preferences, discursive strategies and linguistic realizations, finally leading to the discursive construction of CI.

3.5 Discourse-Historical Approach of Critical Discourse Studies

Our paper adopts the Discourse-Historical Approach (DHA) of Critical Discourse Studies by Ruth Wodak as the theoretical framework for investigating the discursive construction of corporate identities in Chinese and Western CEO statements in annual reports. Developed in the field of discourse studies, the DHA perceives both written and spoken language as a form of social practice (Fairclough and Wodak 1997), which can be seen as constituting non-discursive and discursive social practices and, at the same time, as being constituted by them (Wodak 2001: 66). The DHA distinguishes between three dimensions which constitute textual meanings and structures: the theme contents or topics of specific discourses, discursive strategies and linguistic means that are drawn upon to realize both topics and strategies (Wodak 2001: 72). The DHA proposes five discursive strategies, namely, nomination, predication, argumentation, perspectivization, and intensification/mitigation (Wodak 2009: 44). Please see Table 3.

Nomination strategies focus on rhetoric devices like metaphors or membership categorization devices to represent social actors; Predicational strategies may be realized as negative or positive, stereotypical attributes in the linguistic form of implicit or explicit predicates labeled to social actors; Argumentation strategies are the ones through which positive and negative attributions are justified and legitimized; Perspectivization reflects speakers’ involvement and viewpoints in discourse; Intensification/mitigation strategies help qualify and modify the epistemic status of a proposition by sharpening or toning it down (Reisigl and Wodak 2001: 45). All the five strategies can be linguistically realized. Our paper explores the three dimensions: theme contents, discursive strategies and linguistic means in detail by using the corpus tool Wmatrix5 and discusses the three principal research questions mentioned above. Following this analytical path and the research questions, our paper proposes the following stages of conducting the research (see Table 4).

Table 4:

The stages of conducting the research.

4.1 Word Frequency Lists

Table 5 shows a word frequency list of top 20 content words in Corpus CC and Corpus WC, with functional words omitted from the list because they are highly irrelevant to the present study. As seen from Table 5, the first two high frequency content words in Corpus CC are third-person self-reference in the form of “company” (211) and “group” (153). However, the frequencies of the third-person self-reference in the form of “company” and “group” in Corpus WC are only used 28 and 7 respectively. The first two high frequency content words in Corpus WC are first-person plural pronouns in the form of “our” (602) and “we” (457) whereas the frequencies of first-person plural pronouns in the form of “our” and “we” in Corpus CC are 18 and 48 respectively. From the word frequency list of top 20 content words, we can identity different nomination strategies in Corpus CC and Corpus WC. The choice of self-reference can project different identities of the in-groups and out-groups (Tajfel 1981). Chinese CEOs prefer a more detached and inanimate reference (Thomas 1997) to project an objective and authoritative “out-group” identity (Tajfel 1981). By contrast, Western CEOs prefer a more engaging reference to project an affiliated and inclusive “in-group” identity (Tajfel 1981). Examples of full sentences are as follows:

1. 2017 was a year full of challenges for the Company, but the Board, the management and all the staff took new thought, new structure, new momentum, new target as strategic guidance, overcame various difficulties and attributed to the shareholders with sustained and stable growth in operation again for their support and caring for the growth of the Company. (Corpus CC)

2. In 2017, we continued to drive growth in many of our anchor brands, received a record number of product approvals, made significant advances in our R &D pipeline, and further strengthened our hallmark ownership culture. (Corpus WC)

4.2 High Frequency Verbs

Predication strategies can be realized by using implicit or explicit predicates (Wodak 2001, p. 73) which reveal evaluative attributions of negative or positive qualities. From the verb choices, we can identify predication strategies in Corpus CC and Corpus WC. When we examine the verbs, we can know what behaviors the companies have as social agents. Based on these, we can summarize the corporate identities each corpus tries to project. Table 6 shows high frequency verbs in the two corpora, verbs in basic forms (VV0), ING-participle forms (VVG) and ED-participle forms (VVD) respectively. As Table 6 shows, Corpus CC tends to use the verbs “improve”, “plan”, “accelerating”, “enhancing” and “amounted” to demonstrate their corporate ability from the perspectives of rapid development, promising future and great performance. Corpus WC use verbs such as “help”, “continue”, “leading”, “achieved” and “delivered” to demonstrate their corporate ability from the perspectives of corporate achievements and industrial status. By using these predication strategies, Corpus CC tries to construct their CI as a thriving company with a long-term orientation (Hofstede and Minkov 2010) whereas Corpus WC tends to project a CI as a well-established company with leading industrial status. Examples of full sentences are as follows:

1. China will continue to deepen the reform of the medical and health system, accelerate the expansion of high-quality medical resources and the balanced regional layout, build a hierarchical diagnosis and treatment system and provide people with all-round and comprehensive health services. (Corpus CC)

2. The Group continued to increase investment in R & D; the R &D expenses amounted to RMB167.60 million this year, representing an increase of 32.51 % as compared to the corresponding period in 2020. (Corpus CC)

3. I am proud of how our employees have helped us further our purpose and the breakthroughs we achieved in 2021. (Corpus WC)

4. As our business grows and the competitive landscape changes, we are adjusting our operating model to ensure we continue to lead from a position of strength. (Corpus WC)

4.3 High Frequency Adjectives

The extraction of adjectives from Corpus CC and Corpus WC can provide insights into corporate communication and CI the Chinese and Western companies try to promote. Corpus CC has 557 adjectives with a total frequency overall of 18,981 words while Corpus WC has 525 adjectives out of a total of 17,289 words. Both Chinese and Western CEOs like to use adjectives in their statements. Table 7 shows the first 20 adjectives from each corpus.

An examination of the above lists reveals an abundance of adjectives with positive meanings are used in Corpus CC and Corpus WC, which is consistent with the proposition that self-presentation is usually positive (Wodak 2001: 73). Adjectives such as new, scientific, innovative, international or global from both corpora are employed to present an image of modernity and globalization. Adjectives such as environmental, sustainable, continuous or continued from both corpora present an image of having strong environmental awareness and adjectives such as key or significant to present an image of authority in the pharmaceutical industry. All the research objects of the six pharmaceutical companies are modern, international, authoritative and environmentally-friendly. The above adjectives will inspire trust in patients in their fight against new viruses and diseases, and invite or maintain confidence in shareholders in their decisions for future investments.

Distinctive adjectives used by Chinese CEOs such as strategic, various, full and five-year emphasize that in China with a population of over 1.3 billion the healthcare reform needs strategic planning, which is in compliance with the country’s Five-year Plans and also reflects Chinese culture of collectivism and long-term orientation (Hofstede and Minkov 2010). Distinctive adjectives used by Western CEOs such as strong, proud, major and great show that the Western pharmaceutical companies are satisfied with their achievements and the adjectives of bold, potential and regulatory present an image of self-confidence and professionalism.

4.4 Key Semantic Domains

A key domain analysis of the corpus suggests a broad picture of corporate communication and CI the companies seek to define. The key domain clouds and top 20 semantic domains in Corpus CC and Corpus WC are presented in Table 8. The clouds contain domains with statistically higher frequencies when compared to BNC Sampler CG (Spoken) Business and show up to 100 items with log-likelihood (LL > 6.63, p < 0.01) from the corpus. Those in large fonts have particularly high frequencies across the two corpora. In the columns of top 20 semantic domains, O1 is observed frequency in Corpus CC and Corpus WC. O2 is observed frequency in NORMDATA in the reference corpus of BNC Sampler CG Business. % 1 and % 2 values show relative frequencies in the texts. + indicates overuse in O1 relative to O2 while – indicates underuse in O1 relative to O2. When comparing the semantic domains of two related corpora, and only items with log-likelihood (LL > 6.63, p < 0.01) are shown in this study. We presented the top 20 key semantic domains in the descending order of LL value in order to obtain a better picture of the context of the words and the nuances of a message (Murphy 2013; Rayson 2008) and investigated the key semantic domains in combination with the key words in each of the domains and their concordances.

Table 8:

Key domain clouds and top 20 semantic domains.

Corpus CC
Key domain cloud
Top 20 semantic domains
	Item	01	% 1	O2	% 2	LL	LogRatio
1	B3	436	2.30	56	0.04+	1524.92	5.86	Medicines and medical treatment
2	I2.1	457	2.41	427	0.30+	832.40	2.99	Business: Generally
3	T3−	235	1.24	202	0.14+	449.94	3.11	Time: New and young
4	A2.1+	282	1.49	443	0.31+	345.52	2.24	Change
5	Y1	159	0.84	115	0.08+	334.41	3.36	Science and technology in general
6	N1	485	2.56	1437	1.02+	259.71	1.33	Numbers
7	I4	122	0.64	87	0.06+	258.44	3.38	Industry
8	I1.1	216	1.14	353	0.25+	254.83	2.19	Money and pay
9	A1.1.1	545	2.87	1738	1.23+	253.54	1.22	General actions/making
10	T2++	161	0.85	248	0.18+	200.90	2.27	Time: Beginning
11	S7.1+	293	1.54	831	0.59+	169.53	1.39	In power
12	Z5	6293	33.15	39,170	27.75+	164.30	0.26	Grammatical bin
13	S5+	318	1.68	962	0.68+	163.87	1.30	Belonging to a group
14	N3.2+	114	0.60	150	0.11+	163.01	2.50	Size: Big
15	N3.8+	92	0.48	91	0.06+	161.66	2.91	Speed: Fast
16	O2	244	1.29	642	0.45+	159.76	1.50	Objects generally
17	A6.1	34	0.18	4	0.00+	120.45	5.98	Comparing: Similar/different
18	S8+	211	1.11	612	0.43+	117.40	1.36	Helping
19	W5	40	0.21	19	0.01+	101.25	3.97	Green issues
20	X5.1+	51	0.27	43	0.03+	98.74	3.14	Attentive
Corpus WC
Key domain cloud
Top 20 semantic domains
	Item	01	% 1	O2	% 2	LL	LogRatio
1	B3	310	1.79	56	0.04+	1073.19	5.50	Medicines and medical treatment
2	B2−	240	1.39	91	0.06+	695.03	4.43	Disease
3	Y1	142	0.82	115	0.08+	302.27	3.33	Science and technology in general
4	T3−	163	0.94	202	0.14+	267.03	2.72	Time: New and young
5	T2++	177	1.02	248	0.18+	264.26	2.54	Time: Beginning
6	A11.1+	179	1.04	340	0.24+	202.93	2.10	Important
7	N3.2+	119	0.69	150	0.11+	192.56	2.70	Size: Big
8	N1	383	2.22	1437	1.02+	156.06	1.12	Numbers
9	X9.2+	106	0.61	152	0.11+	155.34	2.51	Success
10	N5++	160	0.93	385	0.27+	138.05	1.76	Quantities: many/much
11	B1	87	0.50	136	0.10+	118.60	2.38	Anatomy and physiology
12	S1.2.5+	55	0.32	44	0.03+	117.83	3.35	Tough/strong
13	A2.1+	158	0.91	443	0.31+	109.97	1.54	Change
14	I3.2	29	0.17	3	0.00+	109.27	6.30	Work and employment: Professionalism
15	A1.8+	84	0.49	139	0.10+	108.85	2.30	Inclusion
16	W3	57	0.33	70	0.05+	93.99	2.73	Geographical terms
17	I2.1	140	0.81	427	0.30+	85.14	1.42	Business: Generally
18	X5.1+	44	0.25	43	0.03+	84.28	3.06	Attentive
19	W1	59	0.34	94	0.07+	79.10	2.36	The universe
20	M2	161	0.93	584	0.41+	70.59	1.17	Putting, pulling, pushing, transporting

4.4.1 Common Semantic Domains

After being compared and merged, the common domains are Medicines and medical treatment (B3); Business: Generally (I2.1); Time: New and young (T3−); Change (A2.1+); Science and technology in general (Y1); Numbers (N1); Time: Beginning (T2++); Size: Big (N3.2+) and Attentive (X5.1+). From the key words of the nine common domains in Table 9, we find that the six companies in Corpus CC and Corpus WC are all listed pharmaceutical multinationals, embracing innovation and transformation for new products and treatments and eager to employ science and technology to keep industrial sustainability and provide better healthcare for the world. These characteristics are in line with the conclusion we made when we examined the first 20 adjectives.

4.4.2 Distinctive Semantic Domains

Distinctive semantic domains and key words in each domain in Corpus CC and Corpus WC are listed in Table 10. The differences highlight the preferred communicating topics in Chinese and Western CEOs’ statements in annual reports. After comparing and merging, we find that Chinese and Western CEOs talk about Category A (general and abstract terms), Category I (money and commerce in industry), Category N (numbers and measurement), Category S (social actions, state and processes), and Category W (world and environment) from different angles.

Based on the above key words and close examination of the concordances, we can find preferred communicating topics in Chinese and Western CEO statements in annual reports:

Corpus CC:

1. Category A (general and abstract terms) Chinese CEOs make comparisons for introspection and motivation and prefer to talk about their reform and development strategies in compliance with the 13th Five-year Plan and the 14th Five-year Plan.

2. Category I (money and commerce in industry) Chinese CEOs choose to talk about the challenges, transformations, performances and growth potentials of the healthcare and distribution industry.

3. Category N (numbers and measurement) Chinese CEOs prefer to talk about the deepening of healthcare reform and the acceleration of the research and development process in China.

4. Category S (social actions, state and processes) Chinese CEOs prefer to use nominal words like company, group, chairman, management, board, directors to refer to the organization, stress the authority of the management and express gratitude for the support from the corporate community of shareholders, partners and employees.

5. Category W (world and environment) Chinese CEOs choose to talk about industrial ecosystem, energy conservation, environmental pollution and protection.

According to Hofstede’s cultural dimension theory, these are all typical manifestations of Chinese culture with the characteristics of high-power distance, high uncertainty avoidance, collectivism and long-term orientation. Chinese CEOs prefer the hierarchical and impersonal communication styles by using third-person self-reference to keep distance for respect, to show authoritativeness and credibility and to ensure social stability and political correctness. The collectivism inclines to take more perspective from the organization level and individuals will believe in and rely on organizations. Organizations conform to the country’s Five-year Development Plan and the environmental protection policy, which means long-term orientation and high degree of uncertainty avoidance for the organization. With the official support and endorsement, CEO statements are taken as credible and rational, and organizations win the support of individuals in turn.

Corpus WC:

1. Category A (general and abstract terms) Western CEOs like to stress the great value they have created and the inclusive culture they have built.

2. Category I (money and commerce in industry) Western CEOs prefer to talk about professionalism as healthcare practitioners and show great care for their colleagues.

3. Category N (numbers and measurement) Western CEOs choose to talk about the significant progress they have made in medical science and the number of people they have served around the world.

4. Category S (social actions, state and processes) Western CEOs prefer to stress their strong brand value and their leading position in the pharmaceutical industry.

5. Category W (world and environment) Western CEOs prefer to talk about the global health and environmental problems, their globalization in manufacturing and supply and their world-changing commitments as a global leader of the pharmaceutical industry.

According to Hofstede’s theory of cultural dimensions, these are all typical manifestations of Western culture. Western CEOs are more people-oriented and keep short distance with their putative audience by frequently using first-person plural pronouns. When the topic of work and employment is concerned, they focus on colleagues and practitioners while Chinese CEOs talk more about the whole industry. They talk about Category B (the body and the individual), with the subdomains B2− (diseases, Rank 2) and B1 (anatomy and physiology, Rank 11), regardless of their unpleasantness and negative associations while we find no Category B in the list of the first 20 semantic domains in Corpus CC. Western CEOs are prone to talk about specific diseases or individuals to show more tolerance for unpleasantness and distinctiveness and they also think each individual’s value matters and power should be equally distributed as they are in a culture of low power distance, low uncertainty avoidance, individualism and short-term orientation.