1. Introduction

PhD examination is a unique type of assessment with examiners playing a key role in it (Dobson, 2017). Examiners’ evaluation of a thesis has (at least) a significant impact on the institutional decision as to whether a PhD should be conferred. In many universities of the world, examiners write a report, like journal peer review (Kumar and Stracke, 2018), which is a combination of summative assessment and formative feedback (Holbrook et al, 2014). While some universities (e.g., Australian universities) find the examiners’ reports adequate for the decision on the pass or fail of the dissertation, many universities of the world require that examiners’ reports be complemented with an oral defense session, in which the examination committee (which may or may not include the external examiner) express their evaluative comments and provide the candidate with the opportunity to respond. The dissertation defense sessions (hereafter DDs), also known as vivas in British-affiliated universities, take the form of either closed or open to public ceremonies.

While examiners’ written report is an obligatory and DD or viva is an optional component of PhD examination worldwide (Hartley, 2000), Iranian universities do not require the examiners to write reports; examiners must attend the DDs to verbally express their evaluation. In Iran, a PhD DD is an open ceremony, which functions as complementing PhD dissertation evaluation, and at the same time is a ritual event marking the award of a PhD to the successful candidate, hence a rite de passage to enter his/her chosen community (Izadi, 2013, 2014, 2017a, 2017b; cf also: Swales, 2004; Chen, 2012; Holbrook et al., 2014). Examiners are the key actors in the discourse as they are institutionally expected to express their assessment of the dissertation under question including both positive and negative evaluation of the dissertation and the candidate (Starfield et al., 2017).

The examiners’ talk is an anaphoric reference to a prior entity_ a PhD dissertation. Therefore, it is influenced by the nature and quality of the PhD they are examining. As examiners are assumed to be experts in the topic of the thesis (or at least in some aspects of it), their evaluation of a thesis has implications for the candidate’s potential to enter his/her discourse community, and therefore has a gate-keeping function (Langfeldt and Kyvic, 2011; Starfield et al., 2017). Examiners’ evaluative stance at times involves expressing negative evaluations (cf: Holbrook et al., 2014; Starfield et al., 2017), holding the candidate and their supervisory committee responsible, although it potentially constructively influences the process of revisions. Examiners do not usually follow institutional guidelines and have their own preferences regarding the choice of summative assessment and formative feedback (Holbrook, et al., 2014), which is partly influenced by the quality of the dissertation and partly by their personal style and persona. Examiners’ language constitutes a wide range of speech acts from positive acts of suggestions and offers to the somehow bitter and face threatening acts of criticisms and rebuttals (cf: Izadi, 2013, 2014; 2017a, 2017b). Starfield et al. (2017) report that “examiners’ comments seem to be a personal attack on the candidate, comments seem to be sarcastic, comments seem to construe a very close relationship between the examiner and the candidate, and comments seem to be more about the examiner rather than the thesis, inter alia” (p. 11). Likewise, Izadi (2014) reports on Iranian examiners’ dilemma to balance out the effective delivery of their critical reviews on the one hand, and interpersonal rapport with the candidates and their supervisors, on the other. Examiners’ multiple goals in discourse compel them to take stances that are pertinent to their aspects of professional and personal identity through various roles, acts and linguistic indexes.

There is now promising research on PhD thesis examination, focusing on the examiners’ reports (e.g., Holbrook et al., 2014; Starfield et al., 2017; Kumar and Stracke, 2018), on dissertation defenses in general, mainly exploring the discourse structure of defenses (e.g., Swales 2004, Mežek and Swales, 2016) and on some aspects of Iranian dissertation defenses (Izadi, 2013, 2014, 2017a, 2017b) in particular. While past studies on PhD examination have mainly remained “slight and exploratory at best” (Lovat et al., 2008, p. 67), there is recent interest in the deeper analysis of examiners’ language from linguistics perspectives (Recski, 2005 and Starfield et al., 2017, from Systemic Functional Linguistics; Izadi, 2013, 2014, 2017a, 2017b; Mayahi and Jalilifar, 2022, from Pragmatics perspective), which is in need of expansion. Moreover, research on examiners’ language has focused on examiners’ reports (cf: Starfield et al., 2017), and very few studies have investigated the examiners’ oral expression of evaluation in dissertation defenses from a functional linguistic perspective (but: Recski, 2005; Lau et al., 2021). Expanding the promising recent research in both dissertation defenses and other arenas of oral academic genres (Zare and Tavakkoli, 2017; Kashiha, 2022), we aim to explore this oral academic genre in light of metadiscourse theory. Metadiscourse can be defined as “the linguistic resources used to organize a discourse or the writer’s stance towards either its content or the reader” (Hyland, 2000). Metadiscourse functions to enhance the textuality of the intended meaning while facilitating interaction between speaker or writer and listener or reader. Thus, metadiscourse helps to realize textual metafunction by creating cohesive texts and to realize interpersonal metafunction by creating interactional relationship between participants in the discourse.

Hyland (2004, 2005) argues that metadiscourse acts as a tool to create a relationship between producer and receiver of discourse through the text. He further notes that what adds to the significance of metadiscourse is its role in changing the view of mere information exchange in communication by engaging participants socially. These functions are realized through metadiscourse markers that present participants’ attitudes, personality, and modality. 

Drawing upon a metadiscourse framework developed by Hyland (2005), this study contributes to the growing body of literature by showing what metadiscursive markers constitute Iranian examiners’ language of evaluation, compared with their international counterparts. The results will hopefully unpack an aspect of oral academic review as a specific genre with implications for both examiners and PhD candidates to become aware of the discursive features of the examiner’s language in the context of English for Academic Purposes (hereafter referred to as EAP). The latter is particularly important in that, in Iran, examiners do not and are not institutionally required to write reports, therefore, the whole evaluation is conducted orally in the defense session. Furthermore, for a better visibility and understanding of the EAP review genre in Iran, we compare the Iranian examiners’ style with the international examiners’ style represented in MICASE corpus. In what follows, we review the literature on PhD examination and defense sessions both globally and in Iran. Then we deal with the methodology and theoretical/analytic framework, followed by the analysis and discussion of the two types of datasets in section 4. Finally, we draw our conclusions.

2. Literature review

In this section, we review research on PhD examination and dissertation defenses.

2.1 Research on PhD dissertation examination 

While research on writing PhD dissertation is extant, research on PhD dissertation examination has received considerably less attention and has mainly focused on examination reports. The importance of examiners’ reports is justified in many academic contexts for two reasons. In some universities, like Australian ones, PhD examination is exclusively carried out through the examiners’ written reports (Holbrook et al. 2014). And, in others that require oral defense, examiners’ reports are given a considerably higher weight compared to the oral defense in the evaluation process (Starfield et al., 2017).

Consistent with the role of PhD as training in research, which should prepare the candidate to become a scholar in his/her field, PhD examination generally focuses on the well-established criteria for sound research in each field. However, as the research shows, examiners do not always follow the guidelines of the institution and apply their own criteria for quality research (Mullins and Kiley, 2002). More importantly, even if they follow the institutional criteria, examiners’ evaluation of the same dissertation may vary depending on their depth of knowledge of and engagement in the topic and their ideologies and expectations about a PhD thesis. A PhD dissertation is a joint product of a PhD candidate and at least a supervisor (but normally there are two), although the degree of supervisors’ engagement in the process of research is subject to regional, disciplinary and cultural variation (Paré, et al., 2011). This variability in the supervisors’ engagement impacts on the quality of the thesis, which, in turn, influences the examiners’ reports. Every PhD dissertation is unique, and therefore, variability in PhD assessment can be seen as a norm rather than an exception.

Examiners’ language in the reports has been observed to address both the quality of the research and the potential of the researcher (Holbrook et al., 2014). Therefore, each report is different, depending on the quality of the dissertation under evaluation. For good-quality dissertations, examiners generally personalize their comments and direct them towards the candidate, while for low-quality ones, they direct their comments to the texts (Holbrook et al., 2014). Comparing examiners’ reports in three disciplines, Kumar and Stracke (2018) observe that examiners simultaneously take on assessor (evaluator) and teacher roles by providing summative assessment and formative feedback. They associate summative assessment, which is a mere evaluation of the dissertation, to the assessor role and formative and developmental feedback, which aims at helping the candidate to revise the dissertation, to the teacher role of the examiners. They further suggest that the developmental feedback aspect of examiners’ reports should be highlighted to fulfill the goal of assessment for learning. Expanding on examiner roles, from linguistic perspective, Starfield et al., (2017) analyze examiners’ reports in New Zealand and argue that examiners adopt nine roles of institution, expert, editor, supervisor, peer, evaluator, reporter, commentator, and viva examiner in evaluating a thesis in terms of coherency, range of depth, methodology, standards and publication.   

From pragmatic perspectives, studies report that examiners skillfully balance out the effective delivery of the criticisms and maintaining the interpersonal rapport with the candidates and their supervisors. They achieve this goal by employing various mitigating strategies, including being politically respectful, downgrading their knowledge, self-denigration, personalization and depersonalization of the criticisms (Izadi, 2014, 2017a, 2017b; Lau et al., 2021; Mayahi and Jalilifar, 2022). One should not, of course, rule out the possibility that the criticism-response exchanges might escalate to a heated conflictive argument (cf: Izadi, 2023). Conversely, in Iranian DDs, the review discourse can also be overshadowed by too much care about collegiality; a practice that is open to over-polite evaluations (Izadi, 2016).  

2.2 Research on dissertation defenses

A dissertation defense (DD) refers to an oral examination complementing PhD dissertation examination (Swales, 2004). It functions as an oral examination, a review of the strong and weak points of a dissertation which facilitates teaching and learning, a "rite de passage" for a candidate, who has spent some precious years on a research project, to become a member of his/her professional discourse community (Swale, 2004), and as a place where interactants maintain and consolidate their interpersonal and professional roles and relationships (Izadi, 2017b). It is also a ritual that marks the award of a PhD, featuring a considerable amount of formulaic language. The ritual aspect has been criticized as ‘empty formalities’ (Swales, 2004) in many universities that assign DDs a lesser priority than PhD evaluation. DDs are said to have no significant impact on the evaluation process, and much more weight is given to the examiners’ reports. Opponents also argue that many PhD candidates are informed about the examiners’ decision by their supervisors before the DD is held (Tinkler and Jackson, 2000; Chen, 2012). Swales (2004, p. 169), however, argues, “[DDs] are clearly different in rhetoric, length and character from ritualistic and ceremonial genres of the academy that deal with the awarding of prizes, honors, and degrees…. [however, they] provide an opportunity for an important academic conversation that operates to certify the candidate’s membership in his or her chosen specialization”. Likewise, Dobson (2017) calls dissertation defenses a social practice that is talked into being by the participants.

The world DDs vary in terms of their structure, content, composition of committee members, criteria for selecting external examiners, ceremonial procedures, levels of formality, degree of openness or closeness to public, and length of the sessions not only from one country to another, but also from one university to another and from one discipline to another within the same country (Grimshaw, 1989; Tinkler and Jackson, 2000; Swales, 2004; Recski, 2005; Chen, 2012; Mežek and Swales, 2016). In the UK and many British-oriented universities, a viva voce is an oral examination in a closed room conducted by an external examiner from another institution and an internal examiner, different from the candidate’s supervisor(s). The external examiner is assigned a great deal of power, and whatever decision s/he makes is hard and fast (Swales, 2004). In vivas, generally supervisors are silent if they are invited. In Scandinavia, DDs (or disputas, as in Norway) are held in a large room in the presence of up to fifty audience presided over by a senior university official. They are treated like a serious ceremony, with the participants dressed up, with the “examiners [or “opponents” as they are called in this region] in full academic regalia”, and with the courtroom style of seating arrangement. The length of ‘disputas’ can vary up to a whole day (in Norway), rounded off by an elegant meal. Recski (2005) quotes Fortanet (2003) that Spanish defenses usually take two hours, and there is a strict rule of questioning order. After the candidate has presented his/her thesis for about 45 minutes, the youngest member of the committee starts his/her questions/comments, followed by the seniors up to the president of the “ceremony”, who is the last speaker.

Swales (2004, ch. 5) describes US defense discourse as an open oral examination in the presence of an audience including the candidate’s supervisory committee, dean/head of the department, a representative from graduate studies institute of the university, internal and external examiners, other interested students and even the candidate’s friends and family members (cf: Swales, 2004, for US defenses). The event often lasts about two hours and is typically held in a large colloquium. The committee members include the candidate’s supervisor, his/her thesis reader(s), one or two external examiners, one or two internal examiners and a representative from the Graduate Studies Division of the respective university. Rearranging Grimshaw’s (1989) dissertation defense’ discourse structure, Swales (2004) reports that US defenses comprise four stages: Preliminaries, the defense proper, in camera session, and closing.

3. Methodology

3.1 Data

The study uses two data sets for the analysis. The first set includes the transcription of Iranian examiners’ evaluative discourse (hereafter referred to as Iranian data abbreviated in the tables and examples as `Irn`) in four PhD dissertation defense sessions that took place in four Iranian public universities and represented the discipline of English Language Teaching. The length of each session ranges between 63 to 123 minutes. The transcriptions yield 4 hours and 9 minutes of talk, resulting in 13404 words. The second dataset includes international examiners’ evaluative discourse, that was extracted from four PhD defense sessions from the Michigan Corpus of Academic Spoken English (MICASE: hereafter referred to as International (with capital I) data, abbreviated as ´Int´ in the tables and examples). The International dataset includes transcriptions of four sessions, that range from 57 to 113 minutes in length. The total length is 3 hours and 57 minutes, which comprises 22971 words. Ideally, such a study would benefit from a larger dataset at the level of corpus, but we understand that the generalizability of our findings would be affected by the size of the data.

3.2 Context of the data

Iranian dissertation defenses bear a considerable resemblance to the US defense sessions as described by Swales (2004) in terms of length, openness to the audience, configuration of the jury and discourse structure (Izadi, 2017a, 2017b). Generally, a PhD program in Iran completes with a dissertation (called payan-nameh in Persian, literally meaning the final letter) with the value of minimally eighteen to maximally twenty-four credit-units following successful completion of twelve to eighteen credit-units of courses and a comprehensive exam (Zehtabian, 2009). Evaluation is rendered based on the quality of the form and the content of the dissertation as well as the candidate’s oral defense and his/her ability to discuss the outcome of his/her work. Usually four examiners (at least one from outside the university) are first proposed by the supervisors, and then selected by the graduate council of the department. Examiners are not required and expected to write reports, but to physically attend the defense sessions to express their evaluation.

Defense sessions are chaired by the candidate’s supervisor, who normally has high power relative to others in manipulating the discourse. Supervisors are expected to defend the candidate against the critical comments of the examiners either right after the candidate’s response or after all the examiners have finished their round of Q-A with the candidate (cf: Izadi, 2017a). Therefore, they are considered the secondary audience of the examiners’ talk.

After the candidate’s presentation and the Q-A section, the chair requests the audience to leave the auditorium to let the committee decide on the result of the evaluation. Every member of the committee (except for the representative of graduate studies of the university) marks the dissertation. A form containing the marking criteria is given to the members to be filled in and signed. Then, the chair collects the mark forms and averages the marks (usually by a calculator). The result is the mean of marks given by committee members, which is a number out of 20. Any mark from 14 to 20 is considered ‘pass’, but with different grades. The grades vary from ‘excellent= 18-20’, to ‘very good=16-18’ and ‘good=14-16’.

3.3 Theoretical approach and procedure

An interpersonal model of metadiscourse, developed by Hyland (2005), was used to analyze the data. According to Hyland, metadiscourse is conceptualized as rhetorical and interactive devices that writers/speakers use to organize a discourse and involve the readers in the text. This model consists of two broad classificatory resources: interactional and interactive metadiscourse. Interactional metadiscourse includes linguistic elements that engage the readers in the unfolding discourse through hedges, boosters, self-mentions, and engagement markers.

The interactional metadiscourse categories and sub-categories are as follows:

Interactive metadiscourse includes linguistic devices that guide the readers through a text through transitions, frame markers, endophoric markers, code glosses, and evidentials. The interactive metadiscourse categories and sub-categories are as follows:

3.4 Data analysis procedures

The procedures for the analysis of the data involve the following steps. First, we downloaded the transcriptions of the defense sessions from the MICASE corpus (International dataset) and selected the transcriptions of four PhD defense sessions out of a corpus of video-recorded data collected by the first author between 2010 and 2012 (cf: Izadi, 2013, 2014, for the description of the corpus). Second, we selected the examiners’ language for analysis. Third, we close-read the two datasets to identify, code and count all the lexicogrammatical features that functioned as a metadiscursive device based on Hyland’s (2005) interpersonal model of metadiscourse. In case of any disagreement over the coding of a metadiscourse marker, we carefully examined the context to determine the primary function of the marker in order to avoid the idiosyncrasies which might be imposed by one analyst. This step was demanding, as in some cases, we were required to go through the text a few times to make sure about the validity of the analysis. A challenging task was to code the lexicogrammatical markers that could ambivalently be attributed to more than one metadiscourse function or the ones that could function as a metadiscourse marker in one context, but not in the other. The word ‘best’, for example, following a discussion, was agreed upon as an attitude marker in such a sentence as ‘it is best to be consistent in reference style’, but not in the phrase ‘to the best of my knowledge’. In the latter case, the word ‘best’ does not actually represent the speakers’ attitude towards a propositional content of a review utterance, but it is part of a phrase that functions as a hedge marker. Next, in order to control the length variation and justify the frequency and distribution of the identified markers across each dataset, we normalized the frequency counts based on Biber et al.’s (1998) raw frequency count/number of words in the text (x 1,000 = normalized frequency count). Normalization is a technique that helps comparison of the results in unequal datasets possible. Finally, the results of data analysis were tabulated and discussed. The markers that enjoyed significant frequency are presented in the tables, but the most frequent ones are selected for discussion. The insignificant markers are all categorized as ‘other’, regardless of their metadiscursive functions, and appear at the end row of the tables.      

4. Results and Discussion

In this section, we present the results of our analysis of the two datasets and discuss the findings. The two datasets were rendered to the analysis of interpersonal model of metadiscourse markers, suggested by Hyland (2005), with its two subcategories of interactional and interactive metadiscourse functions.

4.1 Interactional metadiscourse

Of the interactional metadiscourse markers, engagement markers were found to be the most frequent markers, followed by self-mention markers, hedges and boosters in both datasets. In contrast, the attitude markers had a very low frequency, despite the expectation that the evaluative nature of DDs should produce a significant number of attitude markers. The spoken nature of the genre and the examiners’ willingness to avoid explicit expression of attitudes through overt attitude markers might be the two reasons for this scarce occurrence. This finding is, however, subject to further investigation by studies that specifically focus on the attitude markers.

4.1.1 Engagement markers

Engagement markers help speakers to explicitly step into the discourse by drawing the attention of the listeners to specific issues, concepts, claims or arguments. These functions are linguistically realized by different markers of engagement. Engagement features are important rhetorical resources to anticipate the addressee’s response and to include them in the speaker’s agenda (Hyland, 2005). In creating the spoken genre of review, examiners employ this strategy to engage with their addressees and to co-create the review genre. Engagement markers include two subcategories of questions and directives. They are discursively used by the examiners to involve the PhD candidates in their evaluation. Each of the subcategories can be linguistically realized through interrogative syntax, imperatives, and second person pronouns (you, your).

Table 1. Realizations of engagement markers in PhD defense sessions

The overall high frequency of engagement markers could indicate that the spoken review genre is more of a dialogic mode of academic speech. Engagement markers are linguistic means to realize the interpersonal phase of this evaluative genre, and therefore, help the participants to jointly co-construct the evaluation discourse.   

Breaking down the engagement markers into their subcategories (table 2), we observed the prominence of the second person subject pronoun and possessive adjective in both datasets. Using the second person pronoun ‘you’, the examiners directly address the candidates and involve them in the process of examination. Moreover, even the engaging questions were found to include an element of the second person pronoun/adjective. Another engagement strategy was found to be ‘questions’. Questions were used by the examiners to express their criticisms to the candidates and engage them in the arguments by drawing their attention to the problematic points in the dissertation at hand. This seems to be an effective persuasive strategy when expert examiners in the field strive to recognize the presence of early career researchers and acknowledge their doubts through a set of questions. Understanding the examiners’ questions could consequently result in better understanding of the examiners’ rhetorical and intended meaning. Given these, the main functions of questions were a) to challenge the candidates concerning different parts of their dissertations that the examiners find problematic (Example 1), b) to check understanding (Example 2), and/or to seek agreement and confirmation (Example 3).

Example 1 (Irn): … Is this portfolio assessment?

Example 2 (Int): … Is that the case?

Example 3 (Irn): Okey, so they’re better?

In some cases, engagement markers were used to serve a dual function, when the examiners incorporated directive devices into their questions to ask the candidates to perform actions related to their comment. For example, the examiners asked the candidates to refer to a particular page or table in their dissertations (example 4) or to draw their attention to a particular part of the dissertation (example 5). The frequent engagement markers in both datasets are presented in Table 2.

Example 4 (Int): there’re a bunch, there- there’re many little things in here I’d like you to just take a look at [S2: mhm] but um would you look at sixty-three for a minute?

Example 5 (Irn): Now go to your method chapter as I have serious concerns there.

Table 2. An overview of engagement realizations in PhD defense sessions

As examples 6 and 7 show, directives often lead to a response or reaction from the addresses, and this turns the discourse into dialogic mode. Here the examiners directly impose a question at the end of their comments and anticipate a reaction on the part of the addressees (PhD candidates and their supervisors). Engagement markers are thus necessary resources to guarantee that the candidates are involved in the evaluation process (Examples 6 and 7).

Example 6 (Irn): What’s your problem I mean what is the scientific problem you wanted to solve?

Example 7 (Intl): So what’s your interpretation of um when you’re in, the culture, conflict of psychologists and anthropologists? it’s uh

4.1.2 Self-mention

The results of the distribution of self-mention are presented in Table 3. The table represents a slightly higher frequency of self-mention in the Iranian examiners’ language compared with the international examiners’. The high frequency of self-mention markers in the examiners’ discourse could suggest the dominance of interpersonality level in spoken genres compared with written genres, as reported in a series of studies by Hyland and Jiang (2018), Lee and Casal (2014), and Ho and Li (2018). Since spoken registers are more interactive by nature, speakers are more likely to position themselves in their arguments, mostly through the use of personal pronouns, as a way to signal their authority and stance. Hyland (2004, p. 143) asserts that “academics are able to promote competent scholarly identity and gain credit for their research claims and self-mention has significant consequences for how their message is received”. The use of self-mentions is also a way of personalizing the negative acts and hence mitigating them for relational purposes (Izadi, 2013, 2014).

Table 3. Realizations of self-mention in PhD defense sessions

The results in Table 4 show that self-mention markers were mainly realized by the first-person singular pronoun ‘I’ in the two datasets, followed by ‘me’ in the International (Example 8) and ‘we’ in the Iranian data (Example 9). As argued in the previous research (Vassileva, 1998), the use of first-person pronoun (I) stresses the significance of speaker positioning in spoken discourse. Moreover, the use of the inclusive pronoun ‘we’ could emphasize the supportive role of expert examiners in giving credit to early career researchers and acknowledge their presence by making them realize that they are regarded as a member of the disciplinary community (e. g., As we know, this is due to the disciplinary differences).

Table 4. The most frequent self-mention markers in PhD defense sessions

Such results could be expected, as the examiners consider themselves experts in the field and their self-references are in fact an implicit reference to the expert gatekeepers whose evaluation is critical for the award of the PhD. Therefore, their use of first-person pronoun (I) indexes the voice of the established discourse community members (Examples 8-9). We could further attribute the use of the first-person pronoun (I) to the informal and subjective nature of this discourse.

Example 8 (Int): I’m not quite sure that that actually addresses the question I mean, let me see if i can come up with an example a better, description of...

Example 9 (Irn): We now go to your chapter on page seventy-five and this is one of my main concerns

The results could act as a guide to show that personal pronouns are among the important linguistic features that occupy the subject and other grammatical slots in the clause and perform different functions. Self-mention markers are also means towards the mitigation of negative acts that could endanger the interpersonal relationships between the speakers and the recipients, especially when they co-occur with other hedging devices, as in ‘I would’ (cf: Vassileva, 1998; Izadi, 2014).

4.1.3 Hedges 

Hedge markers are linguistic means to qualify assertions, to indicate uncertainty, and to acknowledge alternative perspectives. Table 5 shows the results of the realizations of hedge markers:

Table 5. Realizations of hedge markers in PhD defense sessions

Consistent with the previous literature (Hyland, 1998, 2005; Koutsantoni, 2006; Muer-Duenãs, 2011, Izadi, 2014, 2017a), hedge markers in this study were among the most frequently utilized metadiscourse devices in PhD defense sessions by both Iranian and International examiners. The high prevalence of hedges in the two datasets could be due the real-time nature of spoken registers in which interlocutors are required to practice more caution in expressing their arguments through the use of, for example, auxiliaries (may, might, can) or adverbs (perhaps, probably). This would consequently prevent them from making strong claims about some assertions. The examiners in both PhD defense sessions intended to “evaluate their assertions in a way that are likely to be acceptable and persuasive” (Hyland, 2004, p. 140) to their audience among whom are discourse participants other than the candidates, as in:

Example 10 (Int): it reminds me I mean I mean, my prediction_ I may be completely wrong

Example 11 (Irn): but probably you thought that the the there were some candidates who took the test simply because they wanted to experience that kind of experience

It could be argued that, by the greater use of hedge markers, examiners appear not only more negotiable but also create a more interactional discourse. The results of the breakdown of hedge markers are presented in Table 6.

Table 6. An overview of hedge markers in PhD defense sessions

The most frequent hedge marker in both datasets was ‘would’, followed by ‘could’ in the International and ‘should’ in the Iranian data. These hedging devices are known as listener-oriented hedging devices, by mitigating assertions through hypothesizing the propositions (examples 12-13).

Example 12 (Int): I mean it, all of those things could be found in early Brubeck recordings where he does these things, [S2: mhm ] uh, uh you know, quite,

Example 13 (Irn): I think it would be a very good idea to be consistent (.)

4.1.4 Boosters

The results of booster markers are displayed in Table 7. As can be seen, boosters enjoyed almost equal frequency in both datasets. This equal occurrence could be attributed to the nature of defense session genre; that is, regardless of linguistic background, signaling certainty and confidence when criticizing the work of PhD candidates is considered the building block of the discourse of examiners. The frequency distribution of boosters in this study is greater than the findings reported by Hyland (1998), Lee and Casal (2014), and Li and Wharton (2012) on written registers. The differences between the findings of this study and those of previous studies, aside from the limited scope of the current data, could also be due to the difference in discourses under investigation. As such, spoken review discourse features criticism and persuasion, which at times requires explicit commitment to the proposition. The gate-keeping role of expert examiners could also suggest that they need to present their arguments with much more authority, expertness, and agency.  

Table 7. Realizations of booster markers in PhD defense sessions

The use of boosters indicated the examiners’ certainty in expressing their epistemic positions and effective evaluation of the dissertation in the defense sessions. The results in Table 8 suggest that ‘know’ was the most frequent booster marker in both sets of data (Examples 14-15).

Example 14 (Int): we know, that, it seems to be at least with respect to some things that Chinese are more, overconfident, than Americans.

Example 15 (Irn): I have worked with these people, so I know that the most problem … 

In Example 15, ‘know’ is used by the examiner to create an interactional relation with audiences by showing certainty about the content of his/her proposition. In example 16, the examiner uses ‘know’ to express certainty or commitment to the proposition (Hyland, 1998) he is asserting, by relying on his first-hand experience.

Table 8. An overview of booster markers in PhD defense sessions

Overall, the international examiners in the PhD defense sessions investigated in this study expressed more engagement markers (64.90/1000W) than their Iranian counterparts (48.59/1000W). Although this finding should await further investigation, using more data, lower frequency of engagement markers in the Iranian data can be attributed to the institutionally relevant practices. In the Iranian universities, where the data were gathered (Izadi, 2017a), the convention was that examiners finished all their evaluative comments, and then the candidates provided their responses to them. This can consequently influence the discourse of the examiners in relation to the number of engagement markers they use. That is, following a pattern of ‘examiner first- candidate second’ by the examiners in these universities would result in less engagement opportunities for interlocuters throughout the discourse of examination.

It was further observed that the International examiners used more hedge markers compared with their Iranian counterparts[1]. This indicates the higher level of certainty of the Iranian examiners in delivering their propositions by using more unmitigated language. Thus, the Iranian examiners preferred not to “acknowledge the provisional nature of their discourse” (Muer-Duenãs, 2011, p. 3073).

4.2 Interactive Metadiscourse

The results concerning the realizations and functions of the sub-categories of interactive metadiscourse markers are presented and discussed in the following sub-sections. Code glosses were found to be the most frequent interactive metadiscourse, followed by transition markers and frame markers.

4.2.1 Code glosses

Code glosses are those linguistic features that are used to clarify the speaker or writer’s intended meaning. They offer more explanations or examples to help facilitate the meaning of previously mentioned propositions (Examples 16-17). Employing such clarification strategies would thus pave the way for smooth and robust interaction.

Example 16 (Int): would you expect, that people’s expectations, would turn out to be the same? so in other words if you were to infer, what they really anticipated, on the basis of their actions, would you, for instance in your experiments, would you expect, that you would in fact get identical expectations as you did? as you did here?

Example 17(Irn): for example like we often said the although they are grammar items but they are different types of grammar items

Table 9. Realizations of code glosses markers in PhD defense sessions

The frequency of code glosses in this study is higher than what is reported in the previous studies (Lee and Casal, 2014; Ho and Li, 2018; Hyland and Jiang, 2018). The differences between the findings could again be related to genre variations; while these studies have focused on written discourse, the current study focuses on spoken discourse. It could be argued that code glosses are more evident in spoken language, as oral registers occur in real time, and as speakers and listeners need more clarifications to create cohesive relations. In written genres, on the other hand, readers have access to the whole text and have the opportunity to navigate through the text a few times to create the necessary cohesive relations between the presented arguments. While writers have frequent opportunities to revise their text before exposing it to the readers, speakers only find the opportunity to revise, repair and reformulate only after their utterances are delivered to their audience.

The most frequent code glosses in both sets of data are presented in Table 10. These markers were used to serve four discourse functions of specifying, qualifying, describing, and extending.

Table 10. The most frequent code glosses markers in PhD defense sessions

4.2.2 Transition Markers

Transition markers function to establish internal relationships between the segments of discourse (Examples 19 and 20). In the current study, transition markers were equally used in the two sets of data (see Table 11). Their equal frequency in the two datasets could indicate that both groups of examiners favored organizing their discourse through a wide range of transition indicators, such as additive (e. g., in addition, moreover), contrastive (e. g., however, by contrast), and resultative (e. g., as a result of, consequently) adverbs. In general, the total number of transitions found in this study was higher than what research on other genres has reported (e. g., Lee and Casal, 2014). For example, transition markers occurred 42 times per 1000 words in this study, while they showed 20 occurrences per 1000 words in Lee and Casal’s (2014) study. Such a difference could suggest that, in a written genre like dissertation, there might be some other discourse markers, apart from transitions, that help the writer to build an internal relationship between parts of the discourse. Another possible justification could be attributed to the nature of genres. Unlike writers, speakers take turns to speak, and at times have to compete for their turn by frequent interruptions of the ongoing speaker throughout the discourse. This would require the speakers to use more transition markers to help maintain the flow of the discourse.    

Example 18 (Int): I was also a bit surprised about the wording, because it seems that…

Example 19 (Irn): I’m saying this because it is becoming a PhD thesis

Table 11. Realizations of transition markers in PhD defense sessions

The most frequent transition markers in both sets of defense sessions were ‘and’ and ‘but’ (Examples 21-22). These items are conjunctions that signal logical relations in a speaker’s thinking and help the listener interpret the possible connections between ideas. They also contribute to the text cohesion and help the audience better interpret the speaker’s intended meaning.

Example 20 (Int): you know where you have two reasonably plausible solutions, and I put him under heavy load, heavy cognitive load, and some time pressure then

Example 21 (Irn): well I’m not paid for (.) the positive points I’m trying to raise but for the negatives ones (.) I will spend more time on the second part

4.2.3 Frame markers

Frame markers signal discourse structures and boundaries such as sequencing, labeling stages, announcing goals and topic shifting in the discourse (Examples 22-23). The results in Table 12 show that examiners intended to frame the structure of their discourse using frame markers. The total number of occurrences of frame markers in this study was far higher than those reported by Hyland and Jiang (2018), Lee and Casal (2014), and Ho and Li (2018). Such discrepancies could again be due to the essence of the discourses under investigation. It could be the case that frame markers are needed more in spoken genres. In written discourse, writers write based on the rhetorical structure of their intended text, so they are less likely to signal the boundaries between discourse parts, while in spoken discourse, speakers need to clearly highlight such boundaries.

Example 22(Int): I first read the um, the abstract was that there’s a class of things called interactionist naive theories, and you’re not gonna tell us which one it is.

Example 23(Irn): we should start with the older, better and more experienced

Table 12. Realizations of frame markers in PhD defense sessions

The most frequent realizations of frame markers in both sets of PhD dissertation sessions are presented in Table 13.

Table 13. The frequent frame markers in PhD defense sessions

The study revealed that both groups of examiners used ‘now’ to label up-coming segment of the discourse, ‘then’ to highlight sequences or stages of their utterance, and ‘well’ to shift the topic, as the following examples show. These functions obviously “assist comprehension and processing” (Hyland & Jiang, 2018, p. 24) on the part of the audiences.

Example 24 (Int): it might, make more sense, to start out right at the very beginning, th- saying what it is you mean, and then get into the content …

Example 25 (Int): you’re not gonna tell us which one it is. but now, as_ after having read it that's not what you meant.

The comparative analysis of interactive metadiscourse in the two sets of data reveal that international examiners used a higher frequency of both code glasses (about four times more than their Iranian counterparts: Table 9), and frame markers (Table 12). Thus, it seems that the international examiners were more interested in conveying a clear message by showing more tendency to produce precise language and to offer a clear account of utterances to their interlocutors compared with Iranian examiners.

5. Conclusion

This paper examined metadiscursive devices in the language of examiners in the dissertation defense sessions that occurred in Iran and the United States. Before drawing any conclusion, we circumspect that, due to the small size of the dataset, the findings are still far from being decisively generalizable. However, we hope to have taken the initial step towards understanding the language of examiners through the analysis of metadiscourse. The discourse that examiners produce in dissertation defenses are constitutive of a specific genre in academic discourse, which could best be termed ‘academic oral review genre’. While in some academic cultures, academic oral review genre may complement the written form of it (as in examiners’ report), in Iran, it is the only form of academic thesis review genre, as examiners are not required to write reports. 

Examiners are the key participants of the dissertation defenses, and their evaluation is central to the award of a PhD to the respective candidate. The employment of a variety of interactive and interactional metadiscursive devices reflects the examiners’ ambivalent desire of effective delivery of their criticisms and suggestions for the improvement of the dissertations at hand, and at the same time, maintaining rapport and positive interpersonal relationships with their audiences, primarily PhD candidates and their supervisors (cf: Izadi, 2013, 2014). This ambivalent desire is well represented, codified and realized through the metadiscursive markers, that are adroitly employed by the examiners. The results of this study unpack an aspect of oral academic review as a specific genre with implications for both EAP-user examiners and PhD candidates to become aware of the discursive features of the examiners’ language. However, as mentioned earlier, the implications of the study and the generalizations of the findings should be subject to some reservations.

The results of this study, although tentatively, can offer pedagogical implications for EAP practitioners as well as PhD candidates by drawing their attention to the genre of criticism in general, and the discourse of expert examiners in particular. PhD candidates who are preparing for their defense session need to familiarize themselves with the strategies that examiners employ to foreground discipline-specific evaluations of the dissertation. For example, examiners may frame their discourse or shift the topic through a range of metadiscourse devices in order to directly or indirectly engage with the candidates in the process of evaluation. Awareness of these strategies can consequently result in better preparation for the defense session by the candidates, enabling them to provide more accurate and effective responses to the opposing comments and criticisms. Adjusting to the norms of examiners’ feedback and the goal they pursue through specific linguistic choices would also prepare candidates to build this capacity to receive and accept negative comments on their work without getting disappointed.

The finding that the Iranian examiners use less engagement markers as compared to their international counterparts could implicate that they frame a less dialogic mode of discourse with the candidates. Therefore, it seems pedagogically advisable to inform the potential examiners and committee members about the necessity of the dialogic mode of the interaction.

While this study analyzed and compared the linguistic choices of a limited number of examiners in the dissertation defense sessions that occurred in Iran and the United States, future endeavors can explore such evaluative genres in the dissertation defense sessions not only using a larger data set but also in various academic cultures and contexts. It is also desirable to conduct such investigations across various fields of knowledge to see the possible variations in the ways examiners who belong to different disciplines engage with the candidates to express their evaluations. We also leave it for future research to investigate the cross-cultural differences in the examiners´ language. Our results show that Iranian examiners used less engagement markers and less hedges, less code glosses and less frame markers compared to their international counterparts, which could be due to cultural differences or the Iranian examiners´ lower pragmatic/linguistic competence, since metadiscoursal elements of language always require the highest level of pragmatic competence. 

[1] We leave it for further studies to delve into hedging more deeply to explain if such results are confirmed, and if confirmed, whether such a difference is explainable in terms of the examiners´ pragmatic competence or their culture.