“For brevity, Fig. 5 only presents the results of …”: Visual-Textual Moves of Electronic Engineering Research Articles by Chinese and Native English Speaking Writers

Visuals are essential components in achieving the communicative purpose of research articles (RAs). Besides, the fact that electronic engineering RAs (EERAs) contain a large number of visuals makes it worth investigating these essential components and their rhetorical functions. By adopting the multimodal move analysis framework of Moghaddasi et al. (2019), the present study explored visuals’ rhetorical functions through examination of the visual-textual moves (VTMs) in 24 EERAs composed by Chinese and Native English speaking writers. Comparisons of the visuals and VTMs in the two corpora were carried out for similarities and variations between the two writer groups. The findings showed that first, visuals in the Introduction (I) and the Methods (M) sections demonstrated mainly ontological function, while those in the Results and Discussion (RD) section played more of epistemological and argument functions. Secondly, more similarities, except for some subtle variations, were found between the two writer groups in terms of the employment of VTMs. The findings of VTMs, as well as their similarities and variations between Chinese and native English speaking writers, not only broaden RA genre knowledge for EE researchers and genre practitioners, but also give insights into academic writing instruction and material design for ESP and EAP practitioners.


Introduction
Research article (RA) or journal article has been recognized as a main channel by scholars from various fields to exchange and disseminate their new research findings and achievements.This academic genre, thus, has received much attention from genre scholars due to its worldwide popularity, widespread recognition, easy accessibility and positive influence on the scholars' personal academic growth.With the aim of assisting the large crowd in producing more qualified RAs, genre scholars have kept investigating this indispensable academic genre from various and different perspectives.This is particularly obvious when Swales (1990Swales ( , 2004) ) proposed the CARS models through his analysis of natural and social sciences RA Introductions.From then on, genre analysis of RAs witnessed a huge expansion in scope and methodology.Genre scholars did not only satisfy themselves with genre studies in applied linguistics field (e.g., Sheldon, 2011;Yang & Allison, 2003), but they also enabled their studies to reach to other disciplines such as chemistry (Stoller & Robinson, 2013), law (Tessuto, 2015), management (Lim, 2012) and neurology (Hung et al., 2012), or to add comparative dimension to their investigations (Hirano, 2009;Loi, 2010;Martín, 2003).Not surprisingly, it is noticeable that the majority of genre studies focused on rhetorical structures and linguistic features of a certain language, which Johns (1998) claimed to be because of the fact that the written words are central to genre researchers' disciplinary values and argumentation.On the other hand, other components in RAs such as graphs, charts, and other visual representations were not prominent to their disciplinary values.Thus, they were found to be a less focused topic.However, academic texts nowadays are becoming increasingly multimodal, which generates a necessity of the investigation of the components that also contribute to the communicative purpose of these academic texts besides the textual language.
The current study focuses on visual-textual moves (VTMs), i.e., textual segments associated with the visuals, in the RAs in the electronic engineering discipline (EERAs).
Visuals in RA refer to graphics such as charts, graphs, diagrams, photos, and tables that give considerable support for the arguments and increase the persuasive power of the text.Many authors used visuals to assist their writing, and skilled readers chose to read the visuals first (Huckin, 1987).Furthermore, graphs and tables can give immediate access to data, making the corresponding argument more plausible (Miller, 1998).
Visuals also allow a big quantity of data to be displayed in a little space.Roth et al. (2005) made a point of saying that visual images might be used and should be a subject of ESP study.They also advocated for the use of visual representations to convey data, demonstrate abstract ideas, organize complicated sets of data, ease the integration of new knowledge with current knowledge, improve information retention, mediate thought processes, and improve problem solving.In a nutshell, visuals in academic papers present information or data regarding the validity of the findings, allowing readers to understand how the data is obtained and to interpret the data for themselves.
The function that visuals play in contributing to the overall communicative goals of texts is so important that it should not be overlooked in a genre-based study.
Consequently, it is critical to explore the importance of visuals, how they demonstrate rhetorical functions, and how they interact with textual language in the EERAs.
Among a wide range of academic disciplines that have been investigated by genre analysts, electronic engineering (EE) is a discipline that receives scant attention from genre scholars in the analysis of visuals.This discipline regulates and processes electronic data using such current technologies as computers and electronic devices.Electronic information technology is inextricably linked to or reliant on every aspect of modern society and people's daily lives, making it a cutting-edge field.Technology has made our lives much easier, allowing us to live more conveniently and work more effectively.The knowledge of VTMs would ease the writing process by not only raising EE researchers awareness of the visuals' functions but also bringing some insights on the effective integration of the visuals into the EERAs, thus fully and accurately contributing to the communicative purpose of each section of the EERAs.Furthermore, examining the VTMs of the two different writer groups could help understand both the common and varied features of the visuals, which offers a feasible and wider choice for these EE researchers for organizing the visuals in their EERAs.Hence, the current study first examines the roles of visuals in the EERAs by examining visuals and VTMs employed by Chinese and native English speaking writers.Second, the similarities and differences of the visuals and VTMs identified between the two writer groups could mostly reveal features of the visuals and VTMs that these two groups share and do not share.The findings are expected to add to the richness of visual-textual knowledge of the RA genre, while also allow us to learn more about acceptable and viable VTMs within the same field.Thus, the present study addresses two research questions.

RQ1. What are the visual-textual moves in EERAs written by Chinese and Native
English speaking writers?RQ2.What are the variations of the visual-textual moves between the EERAs written by Chinese and Native English speaking writers?Dubois (1980) researched slides in biomedical presentations, which is generally regarded as one of the first studies of visuals and their meaning-making capacity in academic genres.Later, Johns (1998) conducted a case study on the use of visual representation and visual-textual interactivity (VTI), or the ways in which language and visual representation interact under macroeconomics context.Based on a review of related literature on visual representation in several fields and an exploratory case analysis on one student, she discovered that ESP researchers attempted to focus more on written language than visual representation because written language was what interested them the most and was central to disciplinary values and argumentation, Furthermore, disciplinary values did not prioritize visual representations such as charts, graphs, and tables.Complex visual representations were dealt with in just a small percentage of the investigations.According to Johns (1998), researchers should maintain constant connection with disciplinary specialists in order to advise them about their discipline's values and how they interpreted and employed visual components.Miller (1998), Rowley-Jolivet (2004), O'Halloran (2010), and Morell (2015) used Systemic Functional Linguistics to underpin several visual investigations.Miller (1998) used three metafunctions of systemic linguistics to investigate the visuals in academic and popular journalism, noting that the visuals' capacity to show information in a little space made it a useful tool for arguing claims.Authors, he believed, not only used verbal texts to create their own arguments, but also employed visuals such as figures and tables to display the findings as the argument's foundation.Rowley-Jolivet (2004) undertook a social semiotic study of visual communication in the context of a scientific research genre -the conference presentation.The study implied that the usage of visuals in each field was dominated and impacted by epistemology, approach, and data type.O'Halloran (2010) revealed that each semiotic resource serves a specific purpose: language introduces theoretical notions or a problem; image provides an overview of mathematical participant relationships; mathematical symbol formalizes the relationships.She went on to claim that combining semiotic resources resulted in 'immense semantic expansion' in mathematical texts, which might be more than the total of the three resources' meaning potentials, a point of view that emphasized the importance of visuals in the argument.Some rhetorical studies focused on the interaction between visual and text in describing and demonstrating the theory, so as to enhance the persuasion and knowledge generation of the whole text.Gross and Harmon (2013) found various argument stages where visuals were present, as well as their semiotic implications.They were: 1) iconic (visuals represent the world); 2) symbolic (visuals represent elements of the world); and 3) indexical (visuals indicate causal relationships in the world).There was a focus on the reader's involvement in attributing meaning/function to visuals, which contrasted with Morell's (2015) view that the author assigned meaning to visuals.Hemais (2014) emphasized that texts improved the meanings of theories, but visuals offered a broad meaning of theories that was simpler for readers to comprehend, a useful finding in terms of visuals' functions in RAs.Du at al (2021) investigated the descriptive statements that accompany graphic presentations in scientific research articles, or figure legends, paying particular attention to their keywords and lexical bundles.Their results shed more light on how the discourse of this part-genre is constructed through various choices of single and multiple lexical items, and offer useful pedagogical suggestions to support the writing of this essential component of academic writing.

Related Studies of Visuals in Academic Genre
Visuals also serve as proof for new information (Charles & Ventola, 2002;Morell, 2015).Morell (2015) discovered that the three functions of the visuals she ascribed to nonverbal (that is, visual) materials (NVM) in her analysis of conference presentations overlapped.The decorative flow chart of a process, for example, may be utilized as evidence to support an argument, which is what the expository function does.This conclusion supported the findings of a prior study, i.e.Miller (1998), on the use of visuals in academic text arguments.
The most current study on visual-verbal relations in mathematics was by Moghaddasi et al. (2019).The study linked visuals to textual rhetorical moves, arguing that identifying visual-verbal moves (however, we adjusted the term 'visual-verbal move' to 'visual-textual move' in our present study) at the micro rhetorical level was critical since the data presented by visuals frequently represented new information and were a key component of RA arguments.The researchers investigated visuals that were accompanied by verbal language and those that are not.Visuals referred to by verbal language are connected with the following rhetorical moves: Establishing Presumptions (EP), Announcing Results (AR), Proof (P), and Discussion (D).Visual moves (VMs) were identified when visuals were not connected with verbal moves, that is, when visual moves substitute verbal moves or when verbal language is absent.Three VMs identified by Moghaddasi et al. (2019) were: 1) Defining Visually (DV); 2) Representing the Operation Outcome (ROO); and 3) Embodying the Operation (EO).
Accompanying these four visual-verbal moves and three visual moves, there were three functions discovered in two-thirds of the thirty mathematics RAs: ontological, argumentative, and epistemological.In visuals, the ontological function was used to emphasize assumptions, define concepts, and create complicated mathematical objects.
The argument function in visuals emphasized the visual's role as evidence, clearly supporting a stated assertion; this function was common in discrete mathematics RA visuals.The epistemological role of visuals was primarily used in the revelation of findings and proofs, as well as the assertion of new knowledge.The distinction of the three functions is artificial, because in many situations the images serve several purposes, according to Moghaddasi et al. (2019).Finally, the study proposed a model for multimodal-move analysis, which demonstrated how to use various visual and verbal moves to achieve the goals of RAs, and explained how discrete mathematicians could create and demonstrate new knowledge using the verbal and visual elements in discrete mathematics.Implications of the study indicated that genre practitioners should possibly incorporate the study of visuals in their discourse analysis because visual components, as well as verbal languages, contributed to the completion of text meaning.
Furthermore, the finding of the study may guide young researchers in understanding how to efficiently design and incorporate visuals into arguments in order to learn as well as to improve authorial trustworthiness.
To summarize, prior research on visuals and multimodality in academic genres focused mostly on the role of visual function and visual-verbal interaction in bolstering authors' or speakers' data arguments and increasing the validity of their knowledge claims.Some research on international conference presentations (e.g., Rowley-Jolivet, 2004;Morell, 2015) indicated that visuals served a multifunctional role in bolstering arguments.Science is the major topic of numerous visual studies (Lemke, 1998;Gross, 2007;Graves, 2014;Gross & Harmon, 2013).Findings of these previous studies revealed that application of visuals might vary across disciplines.
Previous academic research on visuals provided in-depth insights into the crucial meaning-making roles visuals play.Among these studies, Moghaddasi et al. (2019) inspired us to conduct VTM analysis in EERAs authored by two distinct writer groups.
The reasons for choosing Moghaddasi et al. (2019) as the reference for the analyses and their framework will be discussed and described in detail in Methodology.

The Corpus
The corpus of our study contains two smaller corpora entitled CH and NS, representing the two different writer groups: Chinese and Native English speaking writers, respectively.The first step of data collection was the determination of the sample size, which was settled on 12 EERAs in each corpus.A disciplinary expert, a Ph.D. candidate in electronic engineering discipline in a Chinese university, was invited to assist in locating EERAs within the scope of journals indexed by SCOPUS for its trustworthy RA quality.This same disciplinary expert was invited to participate in the interview to shed more light on different issues from the analysis results, therefore, he was referred to as the expert informant in the present study.Since the RAs of the two corpora should be in the similar level, we decided to target the journals that might contain both RAs written by these two writer groups.It was decided that 12 journals were considered as the sources from which the RAs of the two corpora come.There were three criteria for the collection of all of our data.First, one prominent feature of the RAs to be regarded as our target data is that they were written by Chinese and Native English writers, which was a step that we need to take to guarantee the authors' nationality and their L1 status.Even though Chinese authors' names seemed straightforward, we also adopted Wood (2001) and Lu and Deng (2019) for checking the authors' L1 status from their profile including their biography, affiliations and bachelor degree institutions.Second, the RAs should be of empirical studies so that they do not suffer from variations caused by structural or organizational differences.This is partly due to the present study being a part of a larger research project that considered organizational structure to be a key parameter that posed a certain impact on the validity of research.Third, the publication years of the RAs were 2019 and 2020 for capturing the recent trend and features of the RAs.

Data Analysis
The analysis procedure of the VTMs includes several steps.The most important and indispensable step is for us to decide the analysis framework of the visuals and visual-textual moves.We decided to base our study on visual rhetoric (Gross & Harmon, 2013) and Moghaddasi et al. (2019) since the former provided a taxonomy of almost all visuals in science, while the latter is one of the most current and very few studies dealing with the visual moves and visual-verbal moves in RA genre.They both serve as theories as well as inspirations of the present study, with the latter also proposing a model for multimodal-move analysis.The use of Moghaddasi et al.'s (2019) paradigm for visual move analysis was justified on three levels.First, the framework is created by studying visual-verbal moves and visual moves in discrete mathematics RAs, which was based on how scientists express meanings using visual rhetoric (Gross & Harmon, 2013).Science and engineering are two areas that are intertwined.To a significant degree, visual rhetoric ideas in science might be applied to writing in the engineering area.In the examination of conference presentations, Morell (2015) assigned three roles to nonverbal (i.e., visual) materials (NVM), namely illustrative, decorative, and expository NVM.Expository NVM is more prevalent in sciences and engineering, fulfilling the role of evidence-providing.As a result, visual rhetoric in science and engineering serves the same function, implying that the theories may be used to comprehend visuals in engineering RAs.Second, using a systematic research design and methodology, Moghaddasi et al. (2019) generated some unique insights regarding visual rhetoric in academic writing.To analyze and understand the acquired visual data, it employed a triangulated technique that incorporated textual-visual and informant data, as well as metadisciplinary data in mathematics.With such a thorough and logical study design and implementation, the outcomes and conclusions are legitimate and reliable.
Third, in terms of visual move analysis in RAs, Moghaddasi et al.'s (2019) work is the most current.Few studies in academic genres investigate visuals, which "occupy up to half of a science research article" due to the lack of attention paid to visual components in RAs (Hyland, 2006, p. 53).The majority of studies on multimodal analysis focused on conference presentations, while there are few studies on visuals in RAs.Moghaddasi are typically found near the pointers.As a result, the border of a VTM is defined by the textual language linked with the visual that conveys the same communication goals, regardless of whether or not pointers are used.That is, texts with matching pointers relating to a visual are deemed VTMs, whereas those without pointers but nonetheless related to the visual must be counted as VTMs.One visual might be connected to or explained by several VTMs if it is coupled with one or more sets of texts.
The fourth step is to conduct inter-rater reliability check.We invited another disciplinary expert, who is a lecturer with a doctorate degree in EE discipline in another Chinese university, to participate in the inter-rater or coder training of the VTMs and coded the visuals in 15% of 24 EERAs (2 EERAs in both of the CH and TH corpora) with the first researcher of the current study.The results of inter-rater-reliability is satisfying enough to conduct the analysis of all visuals and VTMs in the whole corpus.
The fifth step is to identify and code all of the VTMs of each visual.Thus, two concepts including VTM sequence and VTM pattern were introduced to examine each visual.A VTM sequence is a collection of all the VTMs linked with a certain visual.By taking into account the VTM types, a VTM pattern could be created through a VTM sequence.Assume a single visual has three VTMs connected with it: AR, D, and AR, and the VTM sequence is AR+D+AR, and the VTM pattern is AR+D.
The sixth step was to describe the VTMs, VTM patterns, and offer examples to demonstrate them.Finally, the data from semi-structured interviews were used to triangulate the results.After completing the VTM analysis for the whole EERA, the first researcher asked one Chinese disciplinary informant to engage in a post semistructured interview.The interviewee was a discipline expert who participated in the data collecting process by assisting in the targeting of EERAs.He was a Ph.D. student in EE at a Chinese institution.He was qualified for this interview because of his years of authoring EERAs and publication experience.The interviewee was intended to corroborate the study findings and explain the key findings in order to strengthen the discussion.

Results and Discussion
Since the last section of the EERAs analyzed in this study, the Conclusion, was one section completely without visual, we would hence demonstrate the results and discussion of the I (Introduction), M (Methods), and RD (Results and Discussion) sections of the EERAs.Based on the examination of the 24 EERAs in the two corpora (CH and NS), the visuals were all pointed to by textual segments, a result which was different from that of Moghaddasi et al. (2019) in that some visuals in discrete mathematics were the isolated visual moves which could stand alone to achieve communicative purposes.It is also observed that the visuals in EERAs occurred much more in the Methods and Results and Discussion sections.Overall, there are altogether 159 and 117 visuals in the EERAs of the CH and the NS corpus, respectively; the number of visuals in Methods section of the CH and NS corpus is 69 and 39, respectively; the number of visuals in the RD section of the CH and NS corpus is 89 and 68, respectively.It was found that it is not conventional for EE researchers to include visuals in their Introductions based on the limited number of RAs with visuals in this section (1 in the CH and 4 in the NS).Table 1 offers a summary of the number of the visuals in each section in the EERAs.The fact that visuals occupy a great amount of space in EERAs illustrates their key role in assisting writers with their argument establishment.This feature of EERA is too obvious to be ignored, and it is therefore considered as one conventional characteristic.
The culture of this graphic discipline is found to include a great variety of charts or graphs.
Different from Moghaddasi et al. (2019), visuals in EE field include both the visuals presenting tangible and intangible information.For example, the experimental devices, equipment, site or arrangement of analysis methods might be presented through visuals such as photos, pictures and schematic diagrams, thus participating in the argument for the related claim.The intangible objects in EE field, as those in discrete mathematics (Moghaddasi et al., 2019), are granted with physical presence through visuals.These visuals represent powerful specific concepts and operations within the discipline, bringing readers to the core of reasoning and theories (Angot-Pellissier, 2015).They embody the abstract concepts, complex relations or the products of carefully designed experiments.Visuals in EERAs are a series of observable entities that concretize EERAs' abstract ideas.The visuals identified in the two corpora were 1) line graph, 2) picture, 3) photo, 4)   Examples 1 and 2 illustrate VTM EP.It can be seen from Example 1 that the authors informed readers of the purpose of "compare the uniformity of the radiation under different PV layouts" by employing the proposition "in order to".Then, the authors continued introducing the experimental procedure that achieves that purpose.Example 2 clearly shows the rationale of the employment of certain experimental operation by using the phrase "for …" to reach the purpose of "this" which was mentioned in maximize the EM power coupling from the WR-3 feed to the lens antenna". (1) In order to compare the uniformity of the radiation under different PV layouts, the daily radiation of January 15th and July 15th in Fig. 7 as well as the cumulative weekly radiation in January and July in Fig. 8 are quantified, as shown in Table 5.
For this, the analytical signal-to-noise ratios (ratios of the averaged absorbance values obtained from Fig. 6 to the corresponding absorbance noise values (A/DA)) were compared with the instrumental signal-to-noise ratio (SNR) shown in Fig. 7A.(Figs. 6 and 7,NS9) 2. Announcing results (AR): presenting results is the main communicative purpose of this move.Numerical results accounted for a large proportion of results having been reported in EERAs.However, results could also be presented through visuals, e.g., the accomplished or in-process design or model.The noun "results" is the most apparent signal for this move.Examples 15 and 16 demonstrate VTM AR.
(3) From Fig. 7b, we know that the gain ranges from 0.15 to 4.49 dBic over the whole bandwidth in State II.(Fig. 7, CH12) (4) Figure 3 (B) plots the S0 11 phase seen on the l00 feed for three different lengths.
Phase variations of -13 and +16 compared to ∠S0 22 = ∠S0 33 = 180 can be seen at 60 GHz when the extremes of the search lengths are reached l00 = 2.96λg, l01 = 3.04λg.(Fig. 3, NS8) 3. Discussion (D): usually after presenting the results, authors would provide their explanation or interpretation, which tend to be more subjective than the results.
Examples or evidences for validation of knowledge claims and relating the results to the previous studies and theories are also considered as means of interpretation and explanation.Examples 5 and 6 are illustrations of VTM D. (5) The application of AO systems would therefore be ineffective if the phase grating depicted in Fig. 1 introduces additional significant phase variations due to diffraction.

(NS2) (6)
It is also observed in Figure 6 C,D that with the increase of the frequency, less currents are observed in the arm and increased currents move to the tapered area formed by the bottom part of the dipole where traveling wave distributions are spotted.(CH7) 4. Displaying apparatus or site (DAS): this move introduces devices or components of devices, or it describes the experimental environment.Linguistic features suggesting this move include the phrases such as "consist of", "is made of" as well as verbs such as "present", "show", etc. Numerical parameters are also used for detailed and specific descriptions of the devices and sites.Examples 7 and 8 are demonstrations for VTM DAS. (7) The wafer (NanoLN) consisted of a 500 µm thick silicon handle, a 4.7 µm thick thermal SiO2 layer, and a 600 nm thick x-cut LN thin-film (Fig. 2(a)).(Fig. 2, NS11) (8) The proposed twelve-port MIMO antennas system was fabricated, and its photos taken from the top and back views are presented in Fig. 15   Final dimensions for l00 and l01 were found for the fully integrated antenna system (solid blue curves in Figure 3), yielding values of 2.993λg and 2.976λg, respectively, which are very close to the best performing search combination (dashed black curve in Figure 3).(VTM AR) The stepwise pulses generated by the lC are shown in Fig. 3A (i) as a continuous stream and in Fig. 3B (i) as single pulse which was constructed by the following: 500 steps with 0 V to achieve the base line, 20 steps to achieve 0.9 V, 180 steps to make the pulse top with 0.9 V, 20 steps to bring the pulse signal down to 0 V and the remaining steps to fill the buffer at 0 V (1 step = 10 ns.The pulse generated was repeated with 1 kHz frequency, and the total time duration depends on the number of pulse data that need to be processed.(VTM AR) The corresponding converted currents pulsed from the V-to-I conversion unit, and measured in channel 2 of the lC are shown in Fig. 3 A (ii) and 3B (ii).(VTM AR) IR radiation from the LED was detected by the IR PD and transformed from an optical pulse signal to voltage pulses (Vout) as measured in Channel 1 of the lC shown in Fig. 3A (iii) and 3B (iii).(VTM AR) From Fig. 3 B i) it is observed that the stepwise generated pulse from the mC system follows a smooth shape, with a sharp rise and fall as it is generated.
After detecting the response from the IR LED, the IR PD has a ''rise time" and a ''fall time" of 250 ns a shown in Fig. 3

B. (VTM AR)
A schematic diagram of data processing procedures is shown in Example 10.
Flowcharts, process diagrams, and circuit diagrams, according to Gross and Harmon (2013), are representations in which space is a metaphor for something else, as opposed to representations in which space is space, such as terrestrial maps.In Reading Images, Kress and van Leeuwen (1996) make the distinction between "topographical" and "topological."Similarly, because it primarily represents data processing, the schematic diagram in EERAs of the whole corpus fits into this category.In contrast to a line graph, which depicts real-time trends or data patterns, schematic diagrams, flowcharts, process diagrams, and circuit diagrams are "models of events that just happen to take place in time," which implies that the "sequence" is all that matters.
In Example 10, VTM AR occurred five times since five pieces of texts were found associated with Fig. 3 in NS9.All of them stated the observed facts of a testing action.
The identified VTM sequence AR+AR+AR+AR+AR was categorized into the VTM pattern AR, which was the most prevalent pattern adopted by EE researchers in the two corpora (as shown in Table 5).In the RD section, most visuals, when mentioned by VTM, mainly firstly reported the results or findings, most of which were statistical.
These results showed the obtained facts about a testing of a proposed model or an experiment.

Comparisons of VTMs and VTM patterns
Table 5 shows the summary of all of the four VTMs in the Introduction, Methods, and RD sections of the CH and TH corpora.Since visuals are not essential in the Introductions, the VTMs identified were also quite limited.Based on the small number of visuals appearing in the Introduction section (1 in 1 EERA in the CH corpus, and 10 in 4 EERAs in the NS corpus), the occurrence of the VTMs in this section was also correspondingly confined.i.e., 2 in the CH corpus, and 14 in the NS corpus.Therefore, it is not conventional for EERAs to adopt visuals in their Introductions.The expert informant verifies this finding by explaining "Well, it is not a usual way for us to use visuals in the Introduction.Since we would introduce in detail in the Methods and RD sections those experimental setups and processes.But it also depends on the authors' claim, I mean, if there is a claim in the Introduction for them to make, they could use visuals to help them clarify at the beginning."Therefore, it is similarity rather than variation that was found in adopting visuals between the two corpora.As for VTMs and VTM patterns identified and shown in Tables 5 and 6, we conclude that the visuals in the Introduction section mainly aim to construct a notion or a concept, thus achieving mainly the ontological function.
Tables 5 and 6 demonstrate that VTM EP holds the highest occurrence in the highest number of RAs in the whole corpus.It occurred in 23 EERAs with the occurrence of 110, implying that the visuals' main communicative purpose in the Methods section was to establish presumptions such as informing readers of the experimental rationale, justification, and procedure of their proposed model and design.
Similarity of the adoption of VTM EP between the two corpora was found.The main variation of the VTMs is mainly identified in VTM AR, which occurred in 83.3% of the EERAs in the CH corpus, while it was only employed in 41.7 % of the EERAs in the NS corpus, illustrating that Chinese EE researchers would also adopt visuals to reveal results or findings of certain phrasal experimental procedures.This is less obvious and common for NS researchers.Chinese researchers' preference of EP as well as AR for supporting the visuals in their EERA Methods was interpreted by the expert informant as "We do use visuals to facilitate our claim in visuals and I am surprised that the visuals in the Methods section in Chinese authors' EERAs are that many.I guess this is because the visuals could help those EE researchers clarify and emphasize their hypotheses or assumptions.The findings that visuals also participate in announcing results in the Methods section could be attributed to the research methodology that they adopted and the degree of complexity of the experimental procedure.They might tend to report the results that have impacts on the final or overall findings of their study."the bottom in terms of their total occurrence in each corpus.This indicated that visuals in the RD section of the EERAs mostly exhibited epistemological and argumentative roles, as they were used to present outcomes and conclusions as well as offer accounts and explanation for them.
In addition, noticeable variations were revealed.It was found that Chinese researchers had stronger tendency in adopting visuals to support their arguments based on the number of visuals and VTMs demonstrated in Tables 1 and 5. Second, Table 5 shows "Visuals are key components in the RD section since they assist in displaying the results that is hard to describe only by using textual words.It is usual and conventional to inform our readers of the findings of our study by employing visuals, which functions as evidence for the validity of our findings."The situation of VTM pattern in Table 6 confirms that visuals in the Methods section are mainly for ontological functions.Epistemological function of the visuals in the CH corpus was also revealed from the occurrence of VTMs AR and EP+AR, indicating another variation between the two corpora.

Conclusion and Implication
Visual-textual moves (VTMs) were identified in the Introduction, Methods and On the one hand, visual-textual move analysis could raise ESP/EAP practitioners' awareness of visuals, as they tend to underestimate the importance of visuals.The EERAs' meanings and communicative purposes were achieved through a collaborative effort of textual language and visuals.This is a characteristic of RAs in the science and engineering disciplines.Hence, when ESP and EAP practitioners plan to conduct research in a similar genre, they should pay equal attention to the visuals as they do to the textual language.They may also pay more attention to the visuals than they did previously.On the other hand, visual-textual move analysis could help ESP and EAP instructors understand the connection between texts and visuals.Three rhetorical functions of visuals in EERAs were displayed by the four VTMs identified in this study.
The occurrence of VTM EP implied that the visuals in the Introduction and Methods sections had a strong tendency to demonstrate ontological function, which is in line with the overall communicative goals of these sections.Another example is the visuals in the RD section, which were primarily used to help demonstrate the results and findings and build arguments for them, indicating that the visuals' rhetorical functions-epistemological and argumentative functions-were in line with the overall communicative function of this section.Teachers should consider how the visuals are structured for optimum effect (Miller, 1998).As a result, ESP and EAP instructors might use the findings from this current analysis to teach students how to successfully and effectively integrate visuals into arguments, i.e., how to integrate visuals in the right locations to match the communication function and goal of each part.This is a critical ability for novice researchers in areas that rely on visuals, not just to establish authorial legitimacy, but also to study, understand, and communicate about the topic (Moghaddasi et al., 2019).Additionally, ESP and EAP teachers might strive to teach visual rhetoric to higher-level students by leading them through the process of analyzing visuals and texts, as well as drawing connections between them.This might be accomplished, however, by combining the efforts of ESP or EAP teachers with discipline specialists in the design and execution of academic writing courses.
et al's (2019) work, which represents the most recent trend, serves as a model for the current study on VTMs of the EERAs.The second step of the analysis is to examine the organizational structure of the EERAs in the two corpora, in which we found that not all of the EERAs were organized by Introduction-Methods-Results-Discussion (IMRD) structure.Instead, the structural patterns of EERAs in our corpus are mainly IMRDC, i.e., Introduction-Methods-Results and Discussion-Conclusion.While Introduction and Conclusion sections of the EERAs maintained the general conventional and functional headings, for the body sections, instead of using such general functional headings as "Methods" or "Methodology", EE researchers tended more to employ specific functional headings such as Proposed monitoring system design or Magnetic probe design.This is less true for the Results and Discussion section since EE researchers are found more in favor of the general function heading "Results and Discussion" by combining Results and Discussion sections as one section.The examination and recognition of each section of the EERAs were assisted and guaranteed by the same disciplinary expert mentioned earlier.It was decided afterwards that we would examine the visual-textual moves section by section.The third step is to identify all visuals and their visual-textual moves from 24 EERAs in the whole corpus.The identification of VTM was based on the prominent rhetorical function of each VTM, which largely relies on pointers (e.g., In Fig. 3, Fig 9 indicates that...) of each textual segment.Texts classified as VTMs under examination Tables 2-4 show summaries of visual types in the EERA Introduction, Methods, and Results and Discussion sections, including the number of RAs in which each type of visual occurs and the occurrence of each type of visual.As can be seen from these tables, the dominant visuals in EERAs in the two corpora are line graph, table, picture, photo and schematic diagram.Line graph was identified as the most dominant visual in the EERAs.According toGross and Harmon (2013), line graph functions as representation of data trend.This visual basically consists of horizontal and vertical axes on either two or four side, labels defining the object of measurement, tick marks defining the units of measure; dots, circles or squares symbolizing data points, and solid, dashed, or dotted lines tracing the best fits to data or equations.The perception of the data engages recognition of individual data points through scanning and matching.The result is a proposition in the form of "At point x, the value of y is z." Schematic diagram was found to become one of the dominant visuals that EE researchers employed for illustration of their proposed design or model.Schematic diagram is the topological representation, claiming that space is a metaphor for something else(Gross & Harmon, 2013).The visual types, block diagram and circuit chart, were found to be similar as schematic diagram for regarding space metaphorically.Block diagram is a kind of schematic diagram in which the main functions or parts are represented by blocks, and the wired connections between the blocks indicate the relationship between the blocks.It is often used in hardware design, circuit design, software design, and process flowchart.The function of block diagram is generally to compare the architectural and conceptual description, usually without detailed description of implementation.Circuit chart, another representation for space being a metaphor for the functional relationships that exist among the semiotic components along the virtual circuit starting at one point and ending at another point(Gross & Harmon, 2013), was also identified, however with limited occurrences in the three corpora.In a word, schematic diagram, together with block diagram and circuit chart, metaphorically demonstrates functional relationships that exist among the components(Gross & Harmon, 2013).Contour plot, the last type of newly identified visuals, is to connect the points with the same height of the surface into a loop and directly project them to the plane to form a horizontal curve.However, this type of visual was only identified in 2 EERAs in the NS corpus.The reason that schematic diagram, picture, and line graph are the most popular ones was explained by the expert informant as "Generally, these three kinds of visuals could reflect the main disciplinary philosophy of EE field.The RAs are mainly about the design and development of electronic devices or programs.Thus, these visuals are used to serve as virtual or real evidences of the design that could effectively strengthen the authors' argument, so that these visuals are counted as the most popular ones in our discipline." (a) and (b), respectively.(Fig.15, CH5)The number of VTM EP visuals displayed that visuals in the Introduction and Methods sections were mainly playing Ontological function by introducing materials, instruments, or experimental processes; while visuals in the RD section tended to play both Ontological and Epistemological functions since both EP and AR were identified as the main VTMs in this section.Figure1shows the relationship between VTM and visual.In Figure1, the functions of visuals could be displayed by their VTMs.VTM AR indicates a visual's epistemological function; VTM D indicates a visual's argumentative function; and VTMs EP and D both suggest a visual's ontological function.As the distinction of the functions of visuals is artificial (Moghaddasi et al., 2019), it is very likely that visuals fulfill a multifunctional role, which means that a visual could perform two, or even three functions.There are ample examples showing that a visual could display multiple functions simultaneously.A visual that plays an ontological role by illustrating the process of operation also plays an epistemological role by creating new knowledge through announcing results.This section only shows the overall occurrence or distribution of the VTMs, and it necessary to identify the collection of VTM types of each visual in the whole corpus, which was named as VTM pattern in the present study.Each visual's rhetorical function could be displayed through VTM pattern, which could be obtained from all of the VTMs associated with each visual: namely, VTM sequence.For instance, if a visual is pointed by three VTMs AR, AR, and EP, the VTM sequence of this visual is AR+AR+EP, the VTM pattern should be AR+EP.The number of VTM sequences identified in the whole corpus are 3, 33, and 30 in the Introduction, Methods, and RD sections, respectively.

4. 2 .
Figure 3 plots the magnitude and phase for S0 11 (corresponding to feed l00) as a sample, because this is the case with the most sensitivity to the feed length.(VTM EP)

Figure 3
Figure 3 in NS8 in Example 9 was identified as VTM sequence EP+AR+AR falling that the top VTM was AR in the CH corpus and D in the NS corpus, indicating that Chinese researchers used more visuals when describing their findings than NS researchers.Furthermore, because 42.8% of the 138 VTMs are designated as VTM D, NS authors put greater emphasis to explaining and analyzing the acquired data by offering more comprehensive causes.The fact that 11 RAs in the CH demonstrated VTM D suggested that the majority of Chinese EE writers used visuals to aid in commenting on and giving possible explanations for the results.Our expert informant stressed: The differences in the top three most frequent VTM patterns between the two corpora were evident.There were 24 visuals in the CH that demonstrated the VTM pattern AR+D, accounting for 27% of all visuals.AR+D, on the other hand, was ranked third in the NS for VTM patterns.This revealed that Chinese researchers preferred to employ visuals when presenting findings and offering interpretation.They also employed visuals purely for the purpose of announcing the results.In the NS corpus, situations painted a different picture.Even though VTM patterns D, AR, and AR+D showed no major discrepancy in terms of occurrence, the most often occurring VTM pattern was D, suggesting that NS researchers tended more to involve discussions when referring to the visuals.The findings of the present study, despite multiple functions the visuals play, suggest that visuals in the Introduction and the Methods section perform more of ontological function by establishing concept and notion and by providing justification for experimental setups and procedures.On the contrary, visuals in the RD section mainly tend to play argument and epistemological roles.The findings of the current study, to a large extent, corroborate with that ofMoghaddasi et al. (2019) in the location where these visuals occur in RAs and the main rhetorical functions the visuals perform.That is, ontological function in visuals is usually employed to introduce definitions and operations; argument function in visuals occurred in Introductions or Results and Discussion.Epistemological function in visuals appear mainly in the Results.
Results and Discussion (RD) section by adoptingMoghaddasi et al. (2019) as the analysis framework.The four VTMs existing in the three sections were Establishing presumptions (EP), Announcing results (AR), Discussion (D), which coincided with Moghaddasi et al. (2019).However, Displaying apparatus or site (DAS) was the newly identified VTM in this current study.Different VTMs demonstrate different rhetorical functions of visuals.The three functions, ontology, argument, and epistemology, were obtained from Moghaddasi et al. (2019), too.The comparisons of VTMs and VTM patterns revealed that Chinese and NS writers were similar in adopting the visuals in the EERAs.In the Introduction and the Methods sections, EP was the most dominant VTM in both of the CH and NS corpora, implying that both groups of writers employed visuals in the two sections mainly to introduce certain notation or concept, and purpose and procedure of the experimental action.AR, D, and DAS were the less employed VTMs in both of the CH and NS corpora.What's more, the VTM patterns in the two sections showed the consistent results as the VTMs did.That is, VTM pattern EP was the most dominant one, and AR and AR+D occurred much less in the two corpora.Chinese and NS researchers were more inclined toward VTM and VTM pattern EP, focusing more on only one function, ontological function, of the visuals.In other words, Chinese and NS were more similar in using visuals to perform ontological function, or to assist in informing their readers of their specific methods.The findings of the RD section indicated that Chinese and NS writers demonstrated similarities in adopting VTMs AR and D, which were the top two prevalent VTMs in both of the two corpora.However, a subtle difference suggested that NS writers seemed to lean toward VTM D more than the Chinese counterparts since the visuals in the NS corpus were discovered to be pointed at by both the VTMs AR and D in the similar amount.

Table 1
Summary of the number of visuals in the EERAs

Table 2
Summary of the visuals of the Introduction, Methods, and Results and Discussion sections in the CH and NS corpora Proof (P), was excluded since it was not identified in the two corpora of the EERAs.However, DAS was the newly added VTM in this framework.
As can be seen from Table3, EP is the most frequent VTM in the Introduction, Methods, and RD sections with respective occurrence of 13, 110, 121 in the 24 EERAs, indicating most visuals in these three sections participate in building the notations and concepts for every experimental action.VTM AR, that occurred in 23 Ras with the total occurrence of 104 in the RD section, ranked the second VTM in the RD section, but with a less obvious divergence from the first-ranked VTM EP, which occurred in all of the 24 EERAs with 121 occurrences.Therefore, EP and AR were regarded as the equally important VTMs in the RD section, which is distinct from that of the Introduction and Methods.

Table 3
Summary of each VTM in the Introduction, Methods, and Results and Discussion sections of the whole corpus the authors' claims by presenting the purpose, rationale, and procedure of each experimental action.Related concepts or notations are also introduced in this move.
Table 4 shows the VTM patterns and VTM sequences in the whole corpus.

Table 5
Summary of each VTM in the I, M, and RD sections of the CH and NS corpora

Table 5
also presents that Announcing Results (AR) is the most dominant VTMs in the RD section in both of the CH and NS corpora.It is also apparent that one similarity of AR and D being placed in the top two, while EP and DAS being found at

Table 6
Summary of each VTM pattern in the I, M, and RD sections of the CH and NS corpora