AUDIOSLIDE PRESENTATIONS AS AN APPENDANT GENRE – KEY WORDS, PERSONAL PRONOUNS, STANCE AND ENGAGEMENT

Due to the competitiveness in academic settings and the convenience of technological platforms, several genres have been designed to promote research. The current research studies audioslide presentations, one of the novel academic appendant genres to promote its host genre, research articles. The corpus includes the spoken transcripts collected from oral presentations in the hard science domains and written by expert scholars. We aimed to identify the structure of the presentations, generate keywords used in the talks with reference to an academic spoken English corpus and a daily English speech corpus, and examine how authorial stance and audience engagement are shown in the presentations by employing personal pronouns. The results demonstrate that the speakers tend to highlight the existence of knowledge gaps and the research designs of the investigations, rather than emphasising the findings.


INTRODUCTION
Since academic publication is becoming more obviously competitive in the modern world, promotional features have colonised written and speech events in higher education (Fairclough, 1993;Okamura & Shaw, 2014).Academic texts are now not only informative as traditionally expected, but are also embedded with evaluative and promotional purposes in order to broaden their visibility and readership.Research papers are written not only to deliver new knowledge, but also to project the author's specific contributions in their academic community.Recently, a number of measures have been deployed to achieve promotional purposes with the help of technological devices, thus generating some novel but appendant genre texts in the academic context (e.g.Elsevier, 2014).For instance, some worldleading international journals have gradually required authors to submit separate bullet-pointed highlights which are published in the on-line version, not only to stress the main findings of the research but to arouse potential readers' interest before gaining full access to the paper (Yang, 2016).
In addition to written texts, one of the latest promotion strategies adopted by publishers is to encourage authors to publish webcast-style audioslides1 or audio clips2 to talk about their research on a designated on-line platform.Differing from formal academic conference presentations, there is no author's presence in these audio presentations, only key words and findings which emphasise the relevance of the research together with powerful imagery providing impact, and professional speakers explaining the content (Sellgate, 2015).Audioslides, integrating both written and spoken acts, give authors opportunities to explain their research in their own words, demonstrate their propositions to convince the audience, and promote their work.Audioslides can help readers quickly understand the subject matter and the relevance of the paper (Elsevier, 2016).They are helpful not only for busy professionals but also for the general public who can "hear that there is a real live person behind what can be some pretty abstract ideas" (Gluckman, 2014).In spite of their growing importance, similar to other novel genres in some leading journals, appendant genres such as audioslides are still less explored by language researchers due to their novelty and implicit status (Yang, 2016).Thus, the present study focuses on how authors attempt to draw attention to their accomplishments by constructing effective audioslides, how keywords are deployed to reach the communicative purposes of the audioslides across disciplinary variations, and how researchers present their authorship and acknowledge the presence of the audience in their audioslides, especially through the use of personal pronouns.This could hopefully shed light on the promotional language expressed in this novel and less studied academic spoken genre.A related goal here is also to help research article authors produce their own effective audioslides.

LITERATURE REVIEW
Compared with traditional paper presentations, audioslides have some distinguishing features.Firstly, they are relatively brief in duration, less than five minutes.Secondly, they lack face-to-face interaction with the listeners.Thirdly, the content is of high scholarly assurance as it is based on published and peerreviewed research papers.Hence, audioslides are a peculiar genre embedded with academic spoken acts, apparent promotional purposes and professional discourse.
The evaluative and promotional language in research articles (RAs) has been extensively analysed from many different aspects according to a variety of frameworks such as interpersonality, stance, meta-discourse, etc. Making explicit positive evaluation of one's academic achievements is also becoming very common in the competitive world of academia (Shaw, Kuteeva, & Okamura, 2014).This led to the emergence and heavy use of promotion-oriented 'appendent genres', which are very different from their academic-based host, that is, RAs, in terms of their function and structure.
According to Yang's (2016) classification, two types of genre are found in academic discourse, the host and the appendant genres, depending on their interrelationship.A host genre is a major focus in academic texts, always existing independently.In contrast, an appendant genre can only exist in conjunction with a host and draws less attention, owing to the fact that its space is restricted and that it has a peripheral status (Tse, 2012).The two types of genre have different purposes and exhibit a clear peripheral-nuclear distinction in academic discourse (Shaw, 2016;Yang, 2016).Differing from the host genres, which usually aim to create new knowledge or information, appendant genres are chiefly aimed at sharing information, and are specifically designed to share information relevant to the host genres.They thus often have a much greater promotional purpose.Compared to the investigation on RAs, studies on appendant genres are relatively underrepresented, but they are becoming as prominent as their hosts."Increasingly appendant genres are viewed as indispensable in academic contexts as they perform a higher promotional function to highlight their hosts in the competitive world of academia" (Yang, 2016: 91).These appendant genres such as book blurbs, bio-statements, journal descriptions, highlights and audioslides are thus expected to shoulder the main duties of promoting their host genres, that is, research papers or proposals (Yang, 2016).It is assumed that more innovative appendant genre texts would appear to support their host genres as the rapid development of technology provides an ideal platform for rapidly disseminating those evaluative and promotional messages to the world, surmounting the constraints of time and location (Schwanen & Kwan, 2008).
As noted by Yang (2016) previously, one common strategy used by authors to achieve their promotional and persuasive purposes in academic discourse is to employ certain linguistic devices to represent their stance and engagement with readers, and in so doing connect with them (Hyland, 2005a).As Hyland (2005a) proposes in his stance (a textual voice or community recognised personality following others) and engagement (the writer's relation to readers with respect to the positions advanced in the text) model of studying RAs, these disciplinarysensitive linguistic devices, including hedges, boosters, attitude markers, selfmention, reader pronouns, directives, questions, shared knowledge, and personal asides, are used to express the authors' voices, represent their positions and themselves, and acknowledge their readers' presence (Hyland, 2008a).Projecting authorial stance not only contributes to the interaction between authors and readers, but is also viewed as a way to persuade the audience and promote the researchers' work.Also, studies have underlined not only cross-cultural (Breivega, Dahl, & Fløttum, 2002;Hu & Cao, 2011;Mauranen, 1993;Mur-Dueñas, 2007) but also cross-disciplinary (McGrath & Kuteeva, 2012;North, 2005) differences in the use of linguistic devices in academic rhetoric.Research on how writers present their positions, evaluate arguments, relate to readers and then convince them of propositions in their texts has demonstrated that choices in rhetorical devices and the conventions of using voice and tense in reporting procedures and results as well as in making arguments in different sections of academic texts (Yang, 2015) vary from discipline to discipline (e.g.Hyland, 2001Hyland, , 2005aHyland, , 2005bHyland, , 2008b)).
In contrast to the relatively conventional features of the above language uses, the tendency to use specific or various lexis highlights more clearly the disciplinary variations in academic discourse.One category of these words, that is, keywords, is believed to be important in helping differentiate the textual features.Keywords are words which show a high frequency of occurrence compared to a reference corpus (Bondi & Scott, 2010;Scott, 2014;Scott & Tribble, 2006).Keyword analysis, as a form of qualitative analysis, can help identify the significance of lexis in a given set of academic texts (Groom, 2009) and is beneficial in terms of establishing "a clear understanding of colligational and collocational relationships which generically significant words take on in the discourse" (Tribble, 2013: 137).Analysing keywords has increasingly become of interest in the study of the evaluative features of discourse, and in sketching the lexico-grammatical resources of the texts studied (Martin & White, 2005).Keywords not only indicate the significance and importance of lexis to the texts described (Tribble, 2013), but also often provide "a way of identifying which words best distinguish the texts of a particular author or group of authors from another" (Hyland, 2012: 68).In addition, they are closely associated with the disciplinary cultures, assumptions, and value systems in academic discourse (Groom, 2009).Research on keywords is extensive across various genre types (see Scott, 2014), with a number of studies comparing their use in the soft and hard science domains and demonstrating that differences do in fact exist (e.g.Yang, 2012Yang, , 2013Yang, , 2014)).
However, studies on how speakers represent their stance and relate themselves to their audience by employing certain keywords in the appendant genres, in particular spoken texts, are relatively underrepresented, probably owing to their novelty and marginal status.Yet, the present study argues that promotional culture is quickly intruding into academic written and spoken events, and that the responsibility of promoting academic work no longer lies solely with the publishers, but that authors and researchers are gradually being expected to share the duty of promotion.What researchers have to do now is not only prepare their papers and have them accepted for publication, but also attract more potential readers, gain higher visibility in the academic community, and potentially increase the number of citations to achieve maximum impact (Elsevier, 2014).In other words, researchers are currently being forced to familiarise themselves with these novel appendant genre texts to produce evaluative, interactive, and persuasive texts.Thus, the present study attempts to analyse how one of these genres, audioslides, works as a new academic spoken genre in the hard science domain where authorial-mentions are scarcely used as a convention in academic texts to promote research and interact with the audience through the analysis of personal pronous and keywords.It is hoped that this investigation can bridge the gap in researching academic appendant genres in the literature.
To be specific, the present research aims to answer the following questions: 1. What is the general structure and information included in an audioslide presentation?2. What are the keywords employed in audioslides of the hard sciences which differentiate them from other academic texts? 3. What interactive discourse in terms of self-mentions and reader pronouns are deployed to represent authorial stance and readers' engagement in the audioslides of the hard sciences?

METHODOLOGY
In this section, we describe how the samples of the audioslides were collected from the hard sciences to create the corpus for the present study.Also, the two reference corpora and the analysis instrument used for generating keywords are discussed.

Corpus
All of the audioslides in the present study were collected from a world-leading journal publisher, namely Elsevier, and were taken from the hard sciences.In addition, all of these audioslides are appendant to full papers appearing in SCIindexed journals.The audioslides mainly include the engineering and technology disciplines across different journals such as Engineering Structures, the Journal of Constructional Steel Research or Mechanism and Machine Theory3 .In a competition held by the publisher, only the authors of the most downloaded papers in the engineering and technology journals were invited to produce audioslides (Elsevier, 2013).In total, 67 audioslide sets were created and published on-line for free viewing by the public.In late 2013, the winners of this competition were announced and, interestingly, the three winning presentations based on a mostviewed list were all produced by non-native English authors and speakers, namely Hong Kong Chinese, French/German and Italian.The total number of slides in the 67 audioslide sets was 297, of which 97 were text only, 78 contained only graphic illustrations such as figures, tables, graphs, etc., and the other 122 were composed of both text and graphics.
The collected audioslides were then watched and transcribed.However, due to the different disciplinary backgrounds of the researcher and research assistants, and the research focus on the keywords, evaluative language and interactive discourse in the present study, the audioslides were not transcribed on a word-byword basis but instead only the language which was deployed to represent stance and engagement in the research was transcribed.In other words, disciplinary terminology and specific disciplinary knowledge, where the symbol X was inserted, were not included in the follow-up analysis (see examples of transcription below [1-3]).
(1) Over the next five minutes I will present a standard [project] by Daniel and myself as part of an EU-funded X called X.Our study is about helping in easily recognizable geographic information system-based methodology for assessing the costs a sea level rise in extreme coastal flooding at the local level.
(2) L1 is the bond length and L2 is longer than L1 so that the specimen will fail on the L1 side on the X are loading.In order to control the adhesive X, a special device in figure 3 was designed to make sure all the specimens have identical bone length thickness.
(3) Interestingly, the shape of the X around is mostly estimated, making MT(X) a good feature extractor of motion contrasts.Our goal is X how a richer X percent X given by the diversity.Sometimes X configuration could provide X for high-level task such as actual recondition.
The transcribing task was done by the researcher and his research assistants and was verified for accuracy by one Taiwanese English-speaking colleague.

Instruments and analysis
First of all, to answer research questions 1 and 2, the transcription was imported to the concordancing tool, WordSmith Tools (Scott, 2008).The present major corpus was used as the study corpus, while BNC (British National Corpus) and BASE (British Academic Spoken English) were used as the reference corpora to generate the keywords employed by the speakers and to make a possible comparison between the two corpora.Keywords refer to that lexis which occurs significantly frequently or rarely, compared to the reference corpus, and analysing them can help researchers avoid trivial and insignificant words, and focus instead on what lexis distinguishes the studied texts from other text types (Scott & Tribble, 2006).Keywords contain keyness and aboutness of lexis, showing their particularities in a specific given text (Culpeper, 2009;Tribble, 2013).
Apart from transcribing the audioslides, an observation form was also designed to attend to the semiotic and structural features of the audioslides.The transcribers had to note down what each slide addresses in relation to the research articles (e.g.Introduction, Literature review, Methodology, Results, or Discussion), and the components of each slide (e.g.words only, graphics only, or a mixture).Each audioslide is a very condensed version of a full paper, so only deliberately selected sections are reported in them.This analysis helps to identify which sections of a full paper would be regarded as the most significant by the speaker, and thus why s/he decided to highlight these points to promote the research.Besides, we took a closer look at what keywords demonstrate how the speakers intentionally designed the slides to grab potential readers' attention, to strategically direct them to the relevance of the research, and to successfully ignite their interest in accessing the full paper with a rather short talk and show.The reference corpora used to generate keywords of the present corpus, BASE and BNC, respectively show the relation to academic and ordinary English.
Then, to answer research question 3, to examine the features of interactive language, Hyland's (2005a) model of stance and engagement was adopted with special attention paid to the use of personal pronouns.The wordlist of the corpus generated by WordSmith Tools was used to locate these linguistic devices, and manual coding was applied again.The concordance function of WordSmith was operated to help locate the entries of using the personal pronouns in the corpus.

The structure and information included in audioslide presentations
Table 1 shows the distributions of the slides, representing the IMRD sections individually in each research article.As can be seen, slides containing both words and graphics are the most common design as they help not only convey the core message but summarise the major findings and thus attract the audience's attention.The introduction, method and results sections are the top three sections the presenters addressed.In the hard disciplines, most of the presenters tended to talk more about their motivation for conducting the research due to the lack of previous studies, then proposed novel methods to verify their hypotheses, and provided those of their findings which were confirmed to be significant.A lack of reviewing the relevant literature is natural due to the limited time and purpose of audioslides.However, it is interesting that the presenters did not highlight their findings completely in accordance with the guidelines of designing audioslides offered by the publisher.This is similar to writing 'highlights' in research articles, where researchers may have different interpretations of the actual effects of these promotional appendant genres from the publishers' expectations (Yang, 2016).In academia, 'Publish or Perish' is researchers' top priority, but 'Publish and Promote' could be publishers' major concern.

Keywords employed in the audioslides vs. BASE
Table 2 displays the top 20 over-and under-used keywords in the corpus of audioslides compared to BASE, which indicates its disciplinary peculiarities or differences from a general academic domain.Keywords contain significantly overused and underused (in italics) lexis, compared to the study corpus.The topranking overused keywords, model, descriptors, method, are in line with the previous discussion that hard scientists are inclined to introduce a new model or method in their research design in order to distinguish their research from others.
Optimisation demonstrates a common practice in hard science of evidencing a better method, while tsunami represents the current research needs or academic interest, connected to the huge earthquake in Japan in 2011.The underused keywords, were, quite, much, and might, also explain some conventions of the hard sciences.Past tense, hedges or adverbials are less frequently used in order to present certainty of knowledge.

Keywords employed in the audioslides vs. BNC
Differing from the previous section, Table 3 exhibits the top 20 keywords compared to another reference corpus, BNC, which generates more disciplinespecific keywords.One exception is the overuse of we.Due to the collaborative environment of involving multiple authors in hard science research, the word is vastly deployed to express authorial stance across the talk in the audioslides.In addition, this also implies that, to stress the value of one's research, the use of the first person plural pronoun seems necessary.Hence, it is not surprising to see the words I and me as underused lexis in the list.Another interesting finding is that some overused keywords such as optimisation, modelling, and behaviour were yielded in both corpora.These occurrences confirm the significant keyness and aboutness of the lexis in technology and engineering field, compared to other disciplines and discoursal communities.However, interestingly, the common words a and it are also underused in the corpus.It is assumed that in hard knowledge, specifying the object is more common than using the indicative pronoun, and also repeating the object may be necessary to impress the audience since, unlike when reading articles, listeners have few chances to look back at what the pronoun indicates in a talk.Therefore, explicit indication of objects throughout the presentation can be a strategy for delivering a successful audioslide talk.

The use of I in the audioslides
In the following sections, we discuss how researchers in the hard sciences present themselves in the audioslides presentations by deploying the first person pronouns.In some appendant genres like the current one, hard science researchers are happier than soft science scholars to use the first person pronouns (Okamura & Shaw, 2014;Yang, 2016), which is, however, apparently different from Hyland's (2005a) argument that soft discipline authors tend to project their roles in research, while hard science writers tend to minimise their presence in RAs.This difference evidences that audioslides, as one appendant genre, have a slightly different communicative purpose, compared to its host genre, RAs and therefore, different rhetorical conventions are used to meet those purposes.Some examples were selected for discussion below.
Table 4 shows some features of its use in the hard sciences and Appendix tabulates the frequency and common collocations of all the personal pronouns.Firstly, the tenses can be varied.Not only do present and past tense follow I, but future tense, which is relatively less seen in RAs, is also possible in these slides.Secondly, the presenters used the pronoun across various sections of their papers (IMRD) to claim authorship.They emphasised their presence in the research design and analysis procedure to specifically highlight the unique novelty and absolute significance of their investigations.In line with the conference presentations (CP), presenters in the audioslides also used the first person pronouns, followed by an active verb to "express actions and opinions congruently and take personal responsibility for their decisions and interpretations" (Carter-Thomas & Rowley-Jolivet, 2003: 4).Due to the environment in which much information has to be delivered within the constraints of time, using the first person pronouns in the audioslides can help speakers express their opinions and comments more openly and directly (Rowley-Jolivet & Carter-Thomas, 2005).
Besides, in spite of the fact that all papers in the corpus were multiple authored, some presenters chose I and we interchangeably to represent multiple authorship in the presentations.One possible explanation is that the single speaker decided to take full responsibility for mistakes or errors raised in the presentation.This device is very similar to a common statement in the dissertation/thesis acknowledgements, 'All errors in the dissertation belong to me solely', where the author tries to protect his/her supervisors or committee members.Likewise, here the presenter would perhaps like to protect the rest of the senior authors in their research and decided to solely take the full responsibility for making this presentation.

The use of We in the audioslides
We is more frequently used than I to represent self in the slides, and it is mainly deployed as an exclusive we and as an authorial marker (e.g.we aim to, we would suggest, we propose, etc.) where it only refers to the authors, but sometimes engages listeners as an inclusive we (e.g.we can see [10], as what we can expect [8], we have witnessed [16], and we must know [14]), as exemplified in Table 5.Similar to CP (Rowley-Jolivet & Carter-Thomas, 2005), in the audioslide presentations, the inclusive we signals disciplinary solidarity by engaging the audience in the research choices and decisions, while the exclusive we mainly recounts the actions taken by the team-mates to do research.Different from using I to claim sole responsibility or absolute authorship, exclusive we is the less obvious marker of authority for sharing responsibility and expressing ambiguity (Okamura & Shaw, 2014); besides, it can downplay personal intrusion and yet still stress a unique procedure or important view (Hyland, 2001).In the audioslides distribution of its appearance referring to the different sections in its RA, we in the hard sciences of the slides equally appears in each section of a paper to introduce the topic, to describe the instruments, to discuss the results and to reach the conclusion.In other words, we in hard science presentations is not only limited to addressing methodology or action as it was in Mur-Dueñas' (2007) study.Thus, it seems clear that the use of these pronouns is different because the genres serve different communicative purposes and use two different modes.

N CONCORDANCE
Table 5. Examples of the use of We in the audioslides

The use of Our in the audioslides
Table 6 demonstrates the deployment of the first person plural possessive adjective or determiner, our.The singular possessive my is also found in the corpora but with relatively fewer appearances and with unified usage (e.g.my name is or my presentation) and thus is not discussed.In the hard science talks, our is often followed by the research design, aims and proposals to signify the uniqueness of the investigation.They used our to lead a claim or aim because apparently this authoritative voice is relatively common in the hard disciplines, where research findings are expected to be full of individual innovations or peculiarities, and also knowledge is contributed by group authorship in this appendant genre.However, other than the differentiation of inclusive and exclusive we, interestingly our in the present corpus apparently only refers to group authorship, that is, it is the exclusive our.Speakers used our to highlight the solely possessive contribution or intellectual property of the research team to distinguish their study from previous ones.This usage could be one feature of promoting research in the genre under study.

CONCLUSION
This paper examined a novel but less-attended genre in academic publication, namely audioslides.Authors of journal papers create a 5-minute on-line presentation, talking about and promoting their research to attract the audience's attention.Audioslides from the hard sciences were collected as the study corpus for identifying their structure, generating the keywords and observing the word use for representing authorial stance and engaging with the audience through the use of personal pronouns.The results indicate that, first, the speakers tended to mainly address the motivation and methods of their research in their talks, instead of spending time summarising and highlighting the findings.This inclination is apparently contradictory to the main purpose of creating audioslides as claimed by the publishers (Elsevier, 2016).Secondly, the keyword analysis shows that the hard discipline researchers had many both overused and underused keywords compared to BASE and BNC.The result indicates that there are specific linguistic choices or realisations used as they respond to other communicative purposes associated to this specific genre.Third, in the deployment of personal pronouns representing stance and engagement, the hard science presenters not only tended to use the personal pronouns to claim their authorial stance and engage their audience in the talks, but also to direct the audience's attention to the significance of the research conducted.However, the result also shows that instead of relying on using explicit reader pronouns like you, your, or reader, speakers in the present research tend to engage their audience implicitly while projecting themselves by using the first person plural pronouns.In academic appendant genres, demonstrating how significant their host genre is may be understandable, but it is argued that perhaps speakers could also try to include the audience more explicitly and frequently to create a close connection and mutual rapport as business commercials do if the purpose of the text is to promote a product, either a commercial or academic one.
In addition, the present research has the following significance.First, regarding knowledge, compared to the extensive investigations of interactive language, stance and engagement in written texts, this study focuses on analysing spoken genres in an academic and promotional setting, which is relatively underrepresented in the literature.
Second, due to the development and convenience of technology platforms, the appendant genres embedded with promotional and evaluative purposes, such as highlights, bio-data, video-abstracts and audioslides, are exhibiting their growing importance in the competitive academic arena, and are gradually becoming required by world-leading journals.Researchers are now not only expected to publish their papers but also to promote their research to expand visibility and increase readership.The identification of the keywords in the present study can help authors produce a more attractive and scholastic presentation for their potential audience, thus motivating them to read the full text.In addition, the proper deployment of linguistic devices, that is, the personal pronouns to represent authorial position and engage readers, helps researchers not only make a persuasive proposition but also show academic solidarity with research communities.
Finally, the study also has some pedagogical significance for teaching novice scholars how to design audioslides and make a successful academic audioslide presentation.The results of the study present language teachers and university students, in particular postgraduate students, with the keywords and personal pronouns of an academic speech to persuade the audience and promote research, which would greatly help novice scholars learn how to deliver an effective speech in public and also digitally.
However, since the present research could be the first to explore this new academic spoken genre, several future studies can be carried out to complement it.First, a qualitative design can be integrated in future studies, where researchers can be surveyed about what they decide to report about their research in a very short talk, and how the specific lexis will be used to promote the research.Besides, future research can expand its scope to include audioslides from other disciplines, in particular the soft science domain, to display a more holistic picture of how this novel genre is properly accommodated in the competitive academic environment.In addition, including the talks from both knowledge domains can help identify an established generic structure of audioslide presentations, how this appendant genre is different from its host genres, that is research papers, how the audio relates to the visuals and what the significance of the balance of text and visuals is in the slides.Lastly, other similar appendant genres embedded with promotional purposes in academic publication can be cross-examined and compared to see whether there are similarities or differences in terms of the deployment of lexis and interactive language.

Table 1 .
Layout of the slides and number of slides in each section Vol. 5(1)(2017):

Table 2 .
Top 20 over-and under-used keywords of the audioslides vs. BASE

Table 3 .
Top 20 over-and under-used keywords of the hard science audioslides vs. BNC

Table 4 .
Examples of the use of I in the audioslides