Unsourced evidentiality and critical reading: the case of international postgraduates in Australia

ABSTRACT

This paper investigates readers’ recognition of unsourced evidentials in texts in association with critical reading. To this end, we involved four Iranian postgraduate students at an Australian university in a collective case study where each student read four Persian texts and participated in in-depth interviews and focus group discussions. Data analysis employed critical discourse analysis and discursive power relations to examine participants’ positionalities and subjectivities. Our findings in general confirm the students’ recognition of unsourced evidential in texts as an indication of critical reading. More specifically, it was revealed that linguistic sources (e.g. Persian indefinite suffix ‘−i’), as well as socio-cultural factors such as normalization embedded within the Persian oral narrative and discourse, play significantly in identifying unsourced evidentials. Pedagogical implications for critical reflection on texts containing instances of unsourced evidentials are suggested.

The current study aims to establish potential associations between unsourced evidentials, reading critically (RC), and readers’ awareness of evidential instances embedded in texts. This is in response to the call that the links between linguistic features and power relations embedded in texts need adequate research in critical literacy studies (Janks, 2010, 2018; Luke, 2018). As suggested by Luke (2012), Comber (2015) and Vasquez et al. (2019), readers should be able to critically reflect on texts as this enables them to examine power relations within texts which normally represent inequalities in a society. The current research article also suggests ways to improve readers’ cognizance of unsourced evidential as a clue to the power dynamics embedded in written texts and challenge the inequalities.

Janks (2010, p. 61) believes that RC is a combination of ‘reading with’ and ‘reading against’ texts. Writers endeavor to encourage and position their readers to accept their ideas, and if they agree with almost every idea of the text, this means that they are reading ‘with texts’ (Janks, 2010, p. 96). Readers should be educated to challenge the ideas that are otherwise taken for granted. They should read against texts, which is known as ‘resistant reading’ by Janks (2010, p. 72). Reading against texts is important because texts might serve the interests of certain groups of people, which can subsequently encourage injustice.

There are various approaches to improving RC to direct readers’ attention to inequalities, which for Esquivel (2020) are represented by race and class, for Vetter et al. (2020) through critical conversation, for Weng (2021) by detecting ideologies in EFL contexts, and for Cho and Johnson (2021) by constructing a third space between instructors and students. Whichever of these approaches are taken, the core of the text analysis should be on the workings of linguistic and textual features (Luke, 2018). In this respect, Janks (2010, pp. 74–77) discussed that for the writers to position their readers, they use linguistic devices including ‘lexicalization (word selection for conveying certain purposes), euphemism (for hiding disapprovals or negative implications), and voice (active/passive),’ among others. Zaini (2022b) studied co-text and the manipulation of co-text as a potential way of spreading misinformation. He emphasized that teachers should educate their students about co-text manipulation when practicing RC. Further, Simmons (2018) highlighted the role of modality in RC. These studies remind researchers and educators of the importance of linguistic features in RC and their impacts on readers’ agreement with authors without examining hidden or potential meanings.

Individuals might use unsourced evidential for various reasons such as defending their ideas, lack of adequate information, cultural reasons, or the influence of their education system (Zaini & Shokouhi, 2023). In this paper, we are trying to find out whether these are the only reasons or there might be other sources which could potentially contribute to the associations between RC and unsourced evidential. We initially expound on the definition and various types of evidentiality and review how (un)sourced evidentials have been dominantly addressed in the literature. Then, we unpack how and why our participants were inattentive to the use of unsourced evidentials appearing in texts they read. We also discuss why and how (lack of) awareness of unsourced evidentiality informs RC, and finally offer some pedagogical implications for (critical) literacy classes, and suggest how students’ awareness of the use of unsourced evidentials can be raised.

Evidentiality: from theory to practice

Evidentials are instances or indications that individuals employ to support their ideas, utterances, and statements (Shokouhi et al., 2015). They occur when individuals talk or write to demonstrate their attitudes to knowledge by indicating their degrees of certainty, reliability, and doubt by showing their beliefs, trust, expectations, deduction, induction, inference, and so on (Chafe, 1986). Generally, evidentials are of two types, namely, sourced and unsourced. In the former, the source of information that the speaker or writer mentions is known. For example, in (a) Marie thinks X (mindsay) and (b) John says X (hearsay), the reader/listener knows the source of information. However, in (c) The news was spread last night and (d) Many researchers have already mentioned this problem, which are instances of unsourced evidentials, the source of information is unknown, and the speaker or writer is responsible for the reliability of the information (Shokouhi et al., 2015). Previous studies have examined evidentiality from syntax, semantics, pragmatics, or a combination of these. For instance, Aikhenvald (2015) considers evidentiality as a grammatical category attached to verbs. Hanks (2012) regarded evidentiality as a kind of deixis. Irimia (2018) discussed that evidentiality is more relevant to a syntactic category than a purely pragmatic category.

In terms of semantics, Murray (2021) clarified that two bodies of information are included in evidentials: first is the source of evidence indicating where the information comes from, and the second is the origo or the person who has the information. Further, Roque et al. (2017) investigated evidentials in interrogative contexts cross-linguistically. They found when people communicate, evidentials may flip from speaker to recipient. This means that the semantics of information in evidentials is not merely speaker centric but could be recipient centric. Déchaine et al. (2017) studied evidentiality in association with presenting rather than asserting propositions in common and origo grounds among several languages (e.g., English, Nuu-chah-nulth, etc.). They found that each of these grounds (i.e., common and origo), interlocutors behave differently in different languages. For example, there are differences in terms of illocutionary forces between languages in origo ground. In a different study, Smirnova (2021) in an experimental study scrutinized the associations between evidentials and epistemic modality. The results of her study postulated that there is a nexus between the choice of evidentials and argument strength; that is, the type of argument and its perceived strength determine its use. Then, it is the degree of certainty in arguments, determined by modals, which operationalizes semantic strength and the use of evidentials.

Other studies analyzed evidentiality from a combination of perspectives. Cornillie and Gras (2015) analyzed the semantic, grammatical, and interactional properties of four evidential markers (i.e., evidentemente [evidently], por lo visto [seemingly], al parecer [apparently], and se ve que [it seems]” (p. 141)) in Spanish. They suggested that the use and behavior of these markers vary based on discourse as well as semantic and interactive situations. Moreover, Temmerman (2021) examined the use of evidentials in travel journalism from linguistic and narrative perspectives. She posited that writers use evidentials to make narrative recreation possible and stronger. Bates (2018) analysis of 87 narratives written by patients suffering from eating disorders revealed a nexus between evidentiality and narratives. Narrators tended to use evidential markers to rhetorically communicate their perceptual self when they were experiencing their illness, and their cognitive self, during the recovery period.

Finally, some researchers studied evidentials in educational contexts. Gu (2015) analyzed Japanese school textbooks and found that some conventional evidentials do not appear in history textbooks so long as they are related to Japan’s history of the Second World War. Gu hypothesized that authors use abstract or collective nouns to obfuscate the actor. Miche (2015), through a pragmatic analysis of written discourse, analyzed the use of evidentials in the essays written by French language students and found that three forms of evidentials, namely, ‘the shown source, the quoted source and borrowed knowledge’ (p. 226) are related to the reliability factor and the epistemic information. Yildiz and Turan (2021) compared the use of evidential in doctoral dissertations written by native English speakers and Turkish non-native English speakers. They suggested that the former used more sourced evidential markers when supporting and communicating their ideas. Finally, Bernárdez (2017), emphasizing that evidentiality has been less explored in association with communication and discourse, analyzed its use in communicative and discourse contexts and suggested that the appearance of evidentials in texts depends on the channel (i.e., written or oral language), discourse, and context (e.g., religion).

Although the studies above show that evidentiality has been addressed from various perspectives, the research on the potential relations between unsourced evidentiality and RC is scarce (e.g., Cho & Johnson’s work Cho and Johnson (2021) on critical conversation perspective, Huh’s (2016) on ideology, and Chandrasoma and Ananda’s (2018) on intertextuality). Drawing on Janks (2010), when certain ideologies become prevalent in a society, they appear in texts, and readers normally take them for granted. Despite the merits of the studies above, we do not know whether and how the use of unsourced evidentials inform readers’ interpretations of texts. There is a modicum of research on RC and unsourced evidentiality. Generally, in critical literacy studies, linguistic features and textual analysis have attracted little attention (Simmons, 2018). Hence, the current study raises the following research question:

Whether and in what ways does the use of unsourced evidentials in texts inform reading texts critically?

This study is significant because the associations between evidentials and RC can promote our understanding of the effects of using unsourced evidential in texts. As Janks (2010) put forward, every text tries to encourage readers to see the world through the lenses which the authors introduce. If this is true, certain paradigms of power relations are imposed on readers via linguistic means. If unsourced evidentials influence RC, they can be a part of the range of linguistic tools that Janks (2010) has already introduced. Then, paying conscious attention to the instances of unsourced evidentials could help readers question the ideas embedded in texts before accepting them. This will, consequently, improve individuals’ skills in RC. Finally, as Huang and Wang (2021) suggested, the outburst of the COVID-19 pandemic stresses our need for more sourced information (in relation to mis/disinformation), while everyday texts might not fulfill this requirement.

Conceptual framework

The conceptual framework on which this study is built is weaved by the theories of power of Foucault (1972) and Fairclough’s (2010). As mentioned above, if unsourced evidential conveys power relations, readers should become aware to avoid being ‘ideal readers’ of texts (Janks, 2010, p. 61). To unpack the workings of power relations and establish their associations with unsourced evidentiality, we should start with a definition of discourse. Foucault (1972, p. 107) defines discourse as a set of signs and statements in society which can ‘be assigned particular modalities of existence.’ Discourses are the forces in society that create knowledge, social practices, power dynamics and relations, and subjectivities (Weedon, 1996).

Discourses are important in this study because they are associated with the power-knowledge complex that Foucault (1980) raises. According to Foucault, power is productive or destructive forces within discourses, which allow or disallow the appearance of ideas or objects, videlicet, and the knowledge within them. Knowledge is the product of discourse and forms a complex relationship with power; knowledge cannot be produced without power relations, subsequently reinforcing and maintaining power relations. Second, discourses are associated with subjectivities and identities. Foucault (1982) maintains that there are two forms of subjects related to power relations. One is that individuals are either the subject to others due to control or dependence or are the subject to themselves because of self-knowledge. The other is a form of identity as we understand ourselves. This is associated with discourses and power relations, and it informs our conducts. Power-knowledge complex and subjectivities are appropriate tools in this study as they can demonstrate how the subject to existing discourses in Iranian and Australian societies interprets texts in terms of the use of unsourced evidentials (see Data collection).

To examine how unsourced evidentials appear in texts as objects of certain discourses and how they convey and maintain power relations, we have also employed Fairclough’s (2010) Critical Discourse Analysis (CDA). His model contains three components – texts, discursive practices, and social practices. Social practices form systems of knowledge and belief, identities, and social relations. They are historically-formed practices maintained within societies and, subsequently, are related to discursive practices and power relations existing within discourses. Discursive practices are sets of practices manifested in and reinforced by discourses that create and maintain relationships between individuals through power relations. They are the actual modes of power relations that appear in our societies and demand individuals to act based upon them. The actions emerging from the discursive practices cover a wide range of social acts, such as how individuals should dress, act, use language, etc. As far as this study is concerned, discursive practices within societies, social classes, institutions, and organizations, designate what language (including linguistic features) individuals should use or avoid. To describe text production, Fairclough’s (2010) states that societies are filled with struggles over relations of power and domination as there are constant instabilities between social classes. Consequently, discursive practices and power relations within societies give rise to the production, reproduction, distribution, and transmission of certain language and texts. This means that power relations generally watch what kinds of texts are recurrently appearing, what linguistic features should be used, and what should be avoided.

As an example of the above discussion regarding the use of a particular linguistic feature and its relation to power dominance and evidentiality, as our Persian data have revealed, is the indefinite suffix marker ‘−i’ which plays a great role in this relation. Based on Ibn Rasool and Khoygani (2014, p. 50) on the different roles of this marker, as demonstrated in the following example, the unsourced evidentiality embedded within this marker is used to obfuscate the identity of entities, hence worth investigating further.

(1) Mard-I beougol-Idad

man-indef markertohim/herflower-indef markergave

‘A man gave him/her a flower.’

In example (1), the suffix ‘−i’ is an indication of unsourced evidentiality for two reasons: a) it hides the source of information, and b) it eschews providing enough evidence about the ideas proposed. As Persian ‘−i’ is related to the source of information, we can suggest that it often raises the question of certainty and reliability in association with the source of information. In the example above, since the source of information is not given, the reliability of the information rests on the speaker (Shokouhi et al., 2015).

Methodology

In this section, we explain data collection, participants, and texts. To collect data, we opted for a case study. As RC is associated with an individual’s views and values in examining the hidden messages in a text, a case study is a germane tool for data collection. It seeks to unravel how a phenomenon is informed by individuals’ viewpoints (Denzin & Lincoln, 2011). Accordingly, we involved four international Iranian postgraduates at an Australian university through purposive sampling following the rigorous criteria outlined here. First, the authors and the participants shared Persian language, which could mean that the native language ameliorates fluent communication and expression of ideas (Zaini, 2018).

Second, as related to culture and nationality, we could not involve participants from various backgrounds as an ‘information-rich case’ concept (Robinson, 2014, p. 35) as nonprobability sampling would have been threatened. Moreover, RC has some ‘historical, cultural, and political’ dimensions (Stevens & Bean, 2007, p. 17) which would have been jeopardized if we had involved participants from various backgrounds and cultures.

Third, we chose our participants from both genders as gender in critical literacy is associated with inequalities, power relations, and differences in masculine and feminine worldviews (Keddie, 2008). Fourth, we decided to involve only postgraduates since they need to write assignments and theses, which include a deep level of critical literature review, present rationale arguments, and discussions. Hence, they should demonstrate a deep level of criticality when communicating their ideas. Finally, as the Iranian political system is autocratic, we could not collect such data in Iran. The current regime in Iran does not tolerate individual’s opposing political and religious viewpoints, while the fear of being caught is a running theme (Goldman, 2016). Hence, the researchers and participants would have been at risk when collecting data or publishing subsequent research outputs. Moreover, the participants would not have freely communicated their viewpoints, and it would have jeopardized the trustworthiness of the study. Collecting data in Australia has created another particular avenue because the participants have experienced a real-life educational context in an English-speaking country.

Participants and texts

Participant recruitment was done by distributing online fliers and on campus. Once four participants were employed to participate in this research, we gave them Consent Forms to read and sign. They learned about their rights and responsibilities and were given pseudonyms. Pedram and Soroush, two male participants, completed their bachelors in Iran and moved to Australia to peruse their Master’s in Software and Mechanical Engineering, respectively. Elnaz and Nasrin, two females, completed their Masters in Iran and are doing their PhDs in Australia in Teacher Education and Political Sciences, respectively. Pedram and Elnaz have been in Australia for approximately one year; Nasrin and Soroush have stayed in the host country for almost 9 months. All participants were aged between 29 to 32.

The texts were in Persian and were chosen based on the socio-cultural situation of the society – the complexities of politico-religious challenges in Iran, in the recent score, are to be noted. On the one hand, the economic situation of Iran is experiencing a steep decline, and, as a consequence, various groups protest against the economic and political situations. The president of Iran then, Hassan Rouhani, was being criticized for his sluggish reactions to the economic situation, low wages, unbridled inflation, lack of any clear plans, and even his lackluster nuclear deal with Europe and the U.S. On the other hand, a large group of Iranian women fought for their rights over the compulsory hijab which has eventually led to the current worldwide known slogan ‘Women, Life, and Liberty.’ In such a situation, some opposing discourses against the regime reassure people that the only way to overcome the humiliations created by the current regime is nationalism (see Zia-Ebrahimi, 2016). Given that one of the long-lasting symbols of Iranian nationalism is Ferdowsi – a Persian poet in the 10^th-11^th centuries (Katouzian, 2010) –, two pairs of texts, one in favor of Ferdowsi and one against Ferdowsi, and one text in favor of Rouhani and one against Rouhani were selected.

Data collection

Data collection involved several phases: initially, the participants learned they should reflect on texts using three highlighters, namely, green, red, and yellow. They were asked to read each text and received unlimited time to read them. When reading each text, they highlighted them in green for parts they agreed with, in red for parts they disagreed with, and in yellow for parts they were ambivalent about. The data were collected in the office of the first author of the current study and the whole process took two full working days.

After reading text 1 (the text against Ferdowsi) and highlighting different parts, they were involved in individual interview 1. They were asked on what basis they agreed, disagreed, and were ambivalent about texts. They were asked whether any choice of words or grammar influenced their interpretations of texts. After interview 1, they were involved in a focus group discussion (FGD). In FGD 1, each participant gave their ideas about the text, paragraph by paragraph. After FGD 1, they participated in individual interview 2 and explained whether and why they changed or maintained their ideas about the text.

After a short break, the participants were asked to read the second text, supporting the ideas of Ferdowsi, in the same manner as the first one. In interview 3, similar to interview 1, they clarified how they accepted, rejected, or were ambivalent about text ideas. Once interview 3 was completed, participants were involved in FGD 2 where they discussed their ideas, defended or changed their opinions, and challenged the ideas of others. Then, they were invited to interview 4, where they explained the effects of FGD 2 on their interpretations of the text. Finally, they participated in interview 5 to compare the two texts, explain their ideas about reading opposing texts, using highlighters, and getting involved in FGDs.

The second round of data collection started 2 weeks later, and the same procedures were executed for the second pair of texts: one text in favor of Rouhani’s performance and one text against it. In total, there were forty interviews as well as four FGDs for the 16 times reading of two pairs of opposing texts. Each interview almost took 25 minutes, and each FGD lasted about 90 minutes. In total, we obtained 1360 minutes of raw data which we then coded and transcribed. The design of the study, which involves reading opposing texts, doing FGDs, and interviews, is unique to this project. A team of five researchers contributed to it as follows: first, we discussed several sessions on choosing the texts. Second, given that we needed rich data for the whole project, we created and employed such a design to a) examine the role of reading opposing texts on a topic, b) investigate the role of group discussions on RC, and c) analyze whether and how ambivalence can contribute to RC. Finally, we concluded that reading opposing texts, interviews, and FGDs can have some implications for prospective research projects where future researchers can benefit from this design.

Data analysis

The theoretical background to this study is based on power relations and the ways individuals become subjects to various discourses, which shape their manifold identities and subjectivities, which, in turn, inform how they use language (see Foucault, 1980, 1982). Our theoretical assumptions are, therefore, translated as follows: RC is about the analysis of power relations conveyed via texts (Janks, 2010) while such power relations are rooted in the discourses and discursive practices (Fairclough, 2010). For analyzing our data, however, we employed thematic analysis because it would allow us to clearly categorize our participants’ reactions to the use of unsourced evidentiality to unpack a) how unsourced evidentiality might reinforce categories of power and dominance in society, b) how power relations use instances of unsourced evidentials to maintain their dominance, c) how education systems and discourses within them establish the use of unsourced evidentials, and d) how unsourced evidential becomes normalized.

Then, following Saldana (2016), we first coded the data in several stages, namely, first-cycle coding, second-cycle coding, and subsequent rounds of coding and analysis. We started the first-cycle coding by specifying and chunking parts of participants’ responses. Once it was completed, we started second-cycle coding, which involved adding extra filters and highlights to focus on main and common features. Then, we combined them to form categories by linking recurrent units. Once subcategories were formed, we merged them to create larger categories. One of the categories which formed during this process was unsourced evidentials. The reasons for choosing thematic analysis are that a) it is a highly-flexible method of data analysis (Braun & Clarke, 2006) that helped us conduct coding the data to match with the theories we had adopted, b) it is valuable when researchers aim to include the diversity of ideas from different participants (King, 2004), and c) as a well-structured approach, it helped us summarize and characterize the main features of the study (King, 2004).

As for the reliability of data coding, we met several sessions to arrange the codebook and form the themes. The first author of this study coded the data. The second author, as well as another researcher, then double-checked the raw data and commented on the codebook and its themes. Once a consensus was reached, we proceeded with the translation. Initially, the first author translated the analyzed data into English. Then, the second author read the translated data for accuracy check and provided feedback. Finally, we met on several occasions to revise the translation together.

Finally, to strengthen the reliability of the study, in addition to our positionalities discussed above, including categories of age, we took the following three steps. First, we applied principles of credibility, which is related to reflecting participants’ views by researchers, as accurately as possible (Tobin & Begley, 2004). To meet this, we involved our participants in a series of interviews, FGDs, and reading several texts to collect enough and various data to mirror their views. Then, we used direct quotations to reflect their ideas and recurrent views as accurately as possible. The second principle is transferability. As researchers cannot predict where their findings end up, they should publicize details about the data collection and analysis process, participants, and settings, as much as possible (Creswell, 2009). Third, to meet confirmability, researchers should clearly show how their interpretations are derived from the collected data, why they have chosen certain theories, and how those theories have been implemented in their studies. In this study, we tried to clearly tie the data to theories (Nowell et al., 2017).

Findings

In this section, we present several extractions from our data and analyze them based on the above-mentioned theories. To choose the quotes, we defined and applied a set of criteria as follows: a) all the quotes must contain an instance of unsourced evidentiality; b) quotes have to be taken from different participants; c) they should be obtained from different texts and various stages of data collection (i.e., different interviews and FGDs).

This section is divided into two sub-sections: in the first part, we discuss some instances of unsourced evidentials in texts and participants’ reactions to them. Then, we discuss participants’ use of ‘−i’ as they supported their ideas in the FGDs.

Unsourced evidential in texts

In the interview that followed the text which was criticizing Ferdowsi, Pedram bluntly mentioned that he held pro-Ferdowsi perspectives and criticized the text which was against Ferdowsi:

(2) The text says that ‘many epic poets tell their stories by putting their words in the mouth of animals, birds, objects, etc..’ Then, is the author really comparing Ferdowsi with every epic poet? Ferdowsi is one of the greatest epic poets in the world.

Pedram criticizes the text because he believes that the author’s representation of Ferdowsi is wrong or misleading. However, despite his critical reflection on the text, he did not consider ‘many epic poets’ as an instance of unsourced evidentials and did not question it in the text. Foucault (1982) discusses that subjects are the products of discourses and power relations, and they behave within these frames. As such, Pedram is the subject to academic discourses. He has spent most of his life in Iran. Accordingly, certain references like Ferdowsi are important for him, while other concepts like unsourced evidentials are not (see Zaini & Shokouhi, 2023). Hence, he pays attention to Ferdowsi and disregards ‘many epic poets.’ Soroush, the other participant, after reading the text in favor of Ferdowsi, discussed his ideas in the interview as follows:

(3) I am ambivalent about the following sentence: ‘Some researchers have confirmed that women had active and positive roles in most narrations of The Book of The Kings.’ In fact, the writer could have chosen a better word than ‘positive.’ Probably, ‘influential’ would have been a better word … .

Soroush does not focus on the unsourced evidential, ‘some researchers,’ in his remarks. Rather, he chooses to focus on the word ‘positive.’ He believes the author of the text should have replaced the word ‘positive’ with ‘influential.’ It seems that Soroush likes to use a more specific word, as he thinks ‘positive’ is quite generic and does not fulfill the purpose of women’s role whereas ‘influential’ shows the specific role of women’s positiveness. Paying attention to the level of sophistication of words can be attributed to power-knowledge complex. Drawing on Foucault (1980, p. 69), once knowledge is produced, it ‘functions as a form of power and disseminates the effects of power.’ The role of the subjects is significant as they distribute certain knowledge which subsequently reinforces power relations. The quote above indicates that Soroush is vigilant in his choice of words, attributed to his academic identities and subjectivities. This attention contributes to power-knowledge complex as it maintains certain knowledge (i.e., paying attention to certain words – in this case, ‘positive role’) and reinforces power relations around such knowledge. Relevantly, paying attention to unsourced evidentiality is overlooked. In the FGD following the discussion on the text against Ferdowsi, Nasrin questioned the phrase ‘superficial thinkers’ and said:

(4) This sentence says: ‘Some people, in fact, those shallow thinkers, fools, and ignorant who regard Ferdowsi as a sage and a knowledgeable poet, those groups of people, never accept that Ferdowsi was against women.’ We can criticize individuals because of their ideas but there is no point in labeling them as superficial or fools. The author, instead, could have used different adjectives.

Similarly in this quote, Nasrin focused on the noun phrase ‘shallow thinkers’ but did not read against the unsourced evidentials, which are ‘some people’ and ‘those groups of people.’ The term ‘motefaker-an sade andish’ translated as ‘shallow thinkers’ is a popular term in political debates in Iran, which caught Nasrin’s attention. Drawing on Fairclough’s (2010), texts and the words used in texts, as the final product of discourses, maintain normalized ideas. The use of certain words gradually turns to metaphorization of ‘the transformation of discourse in vocabulary’ (Foucault, 1980, p. 69). As words are used constantly in language, they are historicized in discourses and represent discourses accordingly (Hanks, 2018). In such a situation, the subjects to various discourses either accept those ideas represented through words or resist them through verbal resistance (Foucault, 1980). In the example above, the author labels the proponents of Ferdowsi as ‘shallow thinkers’ – a terminology that is metaphorized in the discourses of Iranian society and represents people who are educated but are not sophisticated (Forouzesh, 2015). Nasrin, as the subject to discourses of academia in Iran, resists this attitude proposed by the author. However, she seems not to pay attention to the instances of unsourced evidential used in the text. Elnaz, in the interview conducted after reading the text against Ferdowsi, assessed the text for spreading religious ideas:

(5) The text says, ‘a group of Ferdowsi worshippers believe that Ferdowsi is a flawless poet without any shortcomings … ..’ Do you know why the author has used the word ‘worship’? To say that fans of Ferdowsi are myopic and prejudiced like those who worshiped idols in Ancient times or those religious people who follow religions without contemplating. I am against these religious ideas.

Elnaz uses her (anti-)religious views to criticize the author. She questioned the use of the word ‘worship,’ which shows her sensitivity toward this word. This can be attributed to the notion of worshiping, as a referential practice (see Hanks, 2018) that catches her attention. That she pays attention to it can also root in the mounting tendency, in Iran, for rejecting Islam, as an inflexible religion. While more people are criticizing the IRI, they are getting more distant from Islam which is being used as a tool to control and oppress the society (see also Boroumand, 2020). Consequently, her sensitivity toward the word ‘worship’ made her overlook the instance of unsourced evidential, ‘a group,’ in the text. The text does not clarify who these worshippers are. After reading the text against Rouhani, Elnaz expressed her disagreement with the word ‘favor’ in the following quote:

(6) I disagree with the use of word ‘favor’ in this sentence: ‘previously, he [the president] has favored some of the representatives of the parliament’ is derogatory because the author is criticizing people, but he is using the word ‘favor’ to indirectly attack the president.

Elnaz, informed by her academic identities, is playing the role of a ‘social actor’ (Fairclough, 2010, p. 75). Drawing on Fairclough’s (2010), there are verbal interactions between Elnaz and text so that Elnaz considers the text and the choice of vocabulary deviant from the norms defined by the discourses with which she identifies. Accordingly, she pays attention to the word ‘favor’ and interprets it as ‘derogatory.’ On the contrary, she does not inquire about ‘some of the representatives,’ which is an instance of unsourced evidential. The next instance of unsourced evidentials was mentioned by Nasrin:

(7) Although I have revisionist attitudes, I disagree with this sentence: ‘The government has had so many achievements since 2013 while, using its whole potential, it has tried its best to serve the people; the government, truly, has provided many stunning services to the people and has an excellent track record, so far.’ I believe the government neither reached nor unleashed its whole potential as specifically women and the younger people were excluded from the cabinet or high-ranking positions, which is not fair. (FGD after the text in favor of Rouhani)

Nasrin, in this example, disagreed with the text based on her political and gender-equality attitudes. This can be attributed to Foucault’s (1980, p. 69) ‘politics of knowledge,’ which is about maintaining knowledge distribution to reinforce forms of ‘domination’ in respect of ‘region and territory.’ Geographical regions, through political systems, maintain certain knowledge to reinforce relations of power and dominance. However, resistance might appear by some groups of people while normalized ideas might remain vibrant for an extended period of time. In the example above, Nasrin by questioning the text for not including women or young people in high political positions, which can be associated with equity, objects to the political system in Iran and wisely reads parts of the text critically. However, there are two instances of unsourced evidentials in the text, namely, ‘many achievements’ and ‘many stunning services.’ Nasrin not only relies on them but also takes them as given. She does not question the text for the instances of unsourced evidentials, which seem to have remained normalized.

Persian indefinite marker ‘−i’

Our data show a recurrent pattern where our participants, in the FGDs, used the indefinite article ‘−i,’ as in the examples below. They are instances of unsourced evidentials that our participants utilized to support their interpretations of texts as they communicated their ideas to their peers. For example, in the FGD, after reading the text against Ferdowsi, Soroush asked the rest of the group why some people accuse Ferdowsi of being anti-women. Then, Pedram said:

(8) dalayel-i hast vali kafi nist

Reasons- indef marker is but adequate not

‘There are some reasons, but they are not adequate.’

In this example, Pedram’s response is an instance of unsourced evidential since he did not mention any reasons in response to what Soroush had asked. Instead, he used the word ‘dalyel-i [some “unknown” reasons],’ which does not clarify the reasons for accusing Ferdowsi of being anti-women. On the one hand, ‘−i’ is the object of discourse in Iranian society. Drawing on Fairclough’s (2010), linguistic features appear in texts due to discursive processes. On the other hand, individuals as the subjects to various discourses employ the available forms, namely, objects of discourse(s), based on the norms of those discourses (Foucault, 1972). Accordingly, Pedram uses ‘−i’ to support his ideas is simultaneously associated with the discourses he is exposed to and the discourses allowing the appearance of ‘−i’ in spoken (or written) texts.

In the FGD, after reading the text in favor of Ferdowsi, Elnaz asked Nasrin to mention some major female characters in the Book of the Kings who have betrayed their countries or their husbands. Nasrin said:

(9) Zanan-i hast-and vali man hozoor zehn na-dar-am

Women-indef marker are but I remember I don’t have

‘There are some women, but I can’t remember.’

In this example, Nasrin did not mention the name of the women, based on Elnaz’s question, but she used ‘−i’ after ‘Zanan [some “unknown” women]’ as an instance of unsourced evidential to answer the question in the FGD. We do not know who those women are and what they have done in the Book of the Kings. This kind of evidentiality marked by ‘−i,’ by Nasrin, takes away the responsibility from her. It is a kind of political response to the question. The use of ‘−i’ as an object of discourse has helped Nasrin maintain the dynamics of power. She has used this particular feature in language to support her ideas, which are related to a domain of knowledge, object, and the subject and contribute to maintaining her status in the group discussion (see Foucault, 1994). In this way, she builds a relation of ‘distance and domination’ (Foucault, 1994, p. 12) to establish her ideas and convince others that she is correct.

In the FGD, after reading the text in favor of Rouhani, Pedram and Soroush both believed that the Islamic Republic of Iran (IRI), as a political system, should be replaced by a democratic political system. While Soroush believed in revisions, Pedram believed in overthrowing the IRI. As they discussed their ideas, Pedram asked Soroush to mention some ways the IRI can adapt or compromise. Soroush mentioned:

(10) Az ye tarigh-i mi-sh-e hal-esh kardFrom one way-indef marker possible solve did‘It can be solved in one way or another.’

This example shows that Soroush used an instance of unsourced evidential ‘ye tarigh-i [one way or another]’ to answer Pedram’s question. Similar to example (9) above, here ‘−i’ shows a political response. However, Pedram who was not satisfied with his answer asked him a further question about how people can make the regime reform itself until it is totally changed to a democracy. Then, Soroush answered:

(11) Khamenei be-mir-e ye kar-i mi-kon-imKhamenei dies one work-indef marker we do‘If the Supreme Leader dies, we will do something.’

These two examples can indicate that individuals as ‘knowing subjects’ employ their knowledge to maintain a ‘domain of objectivity’ (Foucault, 1970, p. 323). This means that the subjects follow an order to maintain power relations associated with the discourses they belong to. What they employ is ‘a language that will be true,’ in relation to given discourses. In other words, discourses determine the truth and the language through which their power relations should be maintained. In example (10), Soroush did not provide any clear answer to Pedram’s query. During the IRI sovereignty, such political, indirect, and irresponsible responses have been given by clerics and authorities to people who have adopted them in their daily language practice. In example (11), likewise, although Soroush was asked to provide evidence, he once again used ‘−i’ to give another unclear response (e.g., ye kar-i mi-kon-im [we will do something]). The use of ‘−i’ is an instance of unsourced information since Soroush does not specify how and through which process they can build a democratic state after the death of the Supreme Leader. Although his comment shows his abhorrence of the IRI as a tyrannical regime, the cycle of giving unsourced evidential via ‘−i’ continues to a point where they stop asking further questions. Hence, Soroush as a knowing subject to discourses of Iranian society constantly uses language that is true (i.e., unsourced evidentials) in the context above. This language follows an order as it can be used repeatedly while it helps the user to maintain the relations of power, as he wins the argument.

Discussion and conclusion

In this section, we discuss the potential roots for participants’ inattention to the use of unsourced evidentiality in texts, explain why they used instances of unsourced evidentials (i.e., ‘−i’) when discussing their ideas, and suggest some pedagogical implications that are related to unsourced evidentiality and RC.

Neglecting unsourced evidentiality in texts

Participants’ inattentiveness to the use of unsourced evidentials can be discussed through the objects produced by discursive practices within cultures. Foucault’s (1977) mentions that discursive practices provide an array of objects. For instance, the objects of geometry, in most cases, are hard objects which can be touched and examined in labs. Objects of discourses can also be linguistic means which have gained ‘conditions of possibilities’ (Janks, 1997, p. 329) – meaning that power relations determine what linguistic choices occur in texts. Unsourced evidentials, accordingly, can be considered as objects of certain discourses (e.g., oral discourses) in the Iranian culture, which gain conditions of possibility to appear in texts.

Pirzadeh (2016) maintains that Persians’ lack of widespread literacy made them resort to oral culture. Governments used to send officials (called Jar’chi) to announce news to the public. The relationships between oral communication and illiteracy, as Pirzadeh highlights, inform the use of unsourced evidentials because oral stories were presented as a monological experience and the source of information from the announcer was never sought. Shokouhi et al. (2015) also suggest that one of the features of oral culture and communication is using unsourced evidentials. Although Iranian society has changed and improved tremendously in terms of literacy, it seems that oral culture is still rampant and relied upon in Iran.

Our participants appear to be influenced by unsourced evidentials as one of the objects of the discourses of oral culture. Given that the use of unsourced evidentials can be attributed to the cultural and historical objects of oral discourses, the order of discourse, social, and discursive practices give rise to the appearance of such devices in texts (Fairclough, 2010). To connect Foucault’s power relation to Pirzadeh’s oral context in the Iranian culture, we maintain that in monologic situations in societies such as Iran with repressive power on voices, the outcome would be consumers’ inability to read or listen to texts non-analytically, hence non-critically. Therefore, the participants did not read against the instances of unsourced evidentials.

Second, unsourced evidentials are considered normal in the discourse of the Persian language by our participants. Shokouhi and Latifi (2019) discuss that in the Iranian education system, schoolbooks are regarded as sources of knowledge, and they cite from Sa’di, a 13^th century great Persian poet, that students are considered as empty vessels that are to be filled with the contents. Students are expected to accept the ideas written in the textbooks, leaving limited room for RC. Given that one of the linguistic features in the discourse of the Persian language is unsourced evidential (Shokouhi et al., 2015), due to the dominance of oral culture, we could suggest that our participants have probably considered unsourced evidential as instances of the normalized discourse of the Persian language.

Normalization, as Foucault’s (1977, p. 177) elaborates, is individuals’ normalized judgment, acceptance, and performance of what is defined by disciplinary power without questioning them. As discussed above, unsourced evidential in Persian can be regarded as norms in that they have already met the conditions of possibility. As Foucault discusses, individuals as the subjects to disciplinary power are required to accept, follow, and perform norms. Normalization is also related to a disciplinary mode of subjectivity. Disciplinary power, as a kind of power whose purpose is to keep people under surveillance, govern their attitudes, behavior, and conducts, demands individuals to improve their performance by adapting themselves to the norms which a given system, institution, or organization defines. The Iranian system of education requires individuals, as its subjects, to accept and consume the ideas used in textbooks (Shokouhi & Latifi, 2019). As such, our participants consumed unsourced evidentials as the subjects to disciplinary power.

Accordingly, this study suggests that instances of unsourced evidentials can appear in texts and position readers since their use might be overlooked by readers. While our participants demonstrated that they can critically read texts related to social, political, national, or gender-associated issues, their analysis of unsourced evidentials in texts remains limited. As for these participants, the historical roots, as well as the education system of Iran (Zaini & Ollerhead, 2019), play crucial roles in terms of their disregard for the use of unsourced evidentiality in texts. However, this question might be raised that our participants have received western education during their stay in Australia. As mentioned above, they have recently migrated to Australia, and the western education they have received might inform their cognizance of the use of unsourced information in texts. Example 10 shows some promise about the effectiveness of the western education that our participants have been receiving.

The use of unsourced evidentiality in FGDs

At the level of production, participants’ use of indefinite article ‘−i,’ an indication of unsourced evidentiality, is significant in supporting their ideas. It is a feature in the discourses of Iranian society and a strategy to circumvent assigning sourced evidentiality. Foucault (1972, p. 64) discusses that different discourses, including the discourse of grammar, generate specific concepts which are formed and regrouped based on ‘their degree of coherence, rigor, and stability.’ Accordingly, ‘−i’ suffixed to nouns can be conceptually important, and contributory to understanding Persian speakers’ critical reading of texts via unsourced evidentiality.

Participants’ use of ‘−i’ can be related to the interplay between objects of discourse and the subjects. As ‘−i’ is an instance of unsourced evidentiality, it is also an object of discourse since it produces and conveys ideas as related to indefinite people, organizations, or entities. Drawing on Foucault (1972, p. 55), discourse ‘is a space of exteriority in which a network of distinct sites is deployed.’ Discourses have systems of formation which create and distribute ideas based on ‘forms of coexistence between statements’ (Foucault, 1972, p. 72). This means that it is impossible for every object to appear in language; however, those authorized to appear are granted permission to come to the surface and used in texts (Fairclough, 2010).

Our data suggest that one of the workings of ‘−i’ is winning arguments due to its attachment to unsourced evidentiality. Examples 8 to 11 above indicate that the participants used ‘−i’ to evade giving direct answers to peers’ questions. This feature which is ostensibly popular not only among politicians but also among ordinary people helps them avoid responding to queries by using general and unspecified comments. For example, Sheikhol Islami (VOA News, 2011), the Minister for Employment during Ahmadinejad’s Presidency (the former President of Iran) mentioned: ‘there are some potentials [Zarfiyatha-i] for creating job opportunities that do not need huge financial resources.’ Similarly, as cited in Pana news Pana (2021), Rezai, a hardliner military commander who was one of the potential presidential candidates in the former election, said in a press conference: ‘I will do something [Kar-i mikonam] that there will be no unemployment.’ In both examples, the politicians skillfully used ‘−i,’ as a technique to evade answering the questions on unemployment. This feature, dominant in the oral culture of Iran, acts as a political tool whenever individuals feel they need to finish an argument by giving oblique answers to questions.

Examples 10 and 11 above also clearly show that Soroush’s response to Pedram’s question using ‘−i’ is an instance of unsourced evidentiality and when Pedram requestioned him, he replied in the same way. This implies that the use of ‘−i’ in marking unsourced evidential is an established marker in Persian discourse and a way to avoid mentioning or blaming authorities directly in their discourse. This is a subtle and clever way that may have its roots in Persian poetry and literary works during the despotic eras (Shokouhi & Zaini, 2022). However, the same cycle can indicate that their propensity for criticality is possible as Pedram questioned the use of unsourced evidential by Soroush.

Hence, we suggest that this study also adds to our understanding of unsourced evidentiality and RC since the findings postulate that readers might use instances of unsourced evidentials when critically reflecting on texts. This highlights that students should be educated to use more reliable information when they support their interpretations of texts. However, we emphasize that the use of unsourced evidentials, when supporting their ideas, is related to these participants; we do not generalize this conduct to others. The oral discourses in Iranian society might have given rise to the use of unsourced evidentiality in support of their ideas. This can indicate that the use of unsourced evidentials as an object of oral discourses is considered ‘legitimate’ (Foucault, 1977, p. 191) by our participants, which, subsequently, underscores the necessity of education about unsourced evidentiality. Moreover, the role of group discussions which can improve critical reflections on texts cannot be disregarded (Shokouhi & Zaini, 2022).

Pedagogical implications

The analysis of unsourced evidentials at the consumption and production levels has pedagogical implications, which can be implemented in various contexts. This means that our case study can offer insights into RC for students and teachers with other language and ethnic backgrounds. While reading texts critically is about examining power relations in texts for the sake of social justice and resisting the dominant ideologies (Janks, 2018), authors of the texts make every effort, including the use of linguistic features, to encourage readers to accept their ideas (Janks, 2010). This hints at the further exploration of the associations between linguistic features and RC (Luke, 2018).

Second, the use of unsourced evidentials might threaten the integrity of our societies as well as social security. As Janks (2010) mentions, societies are filled with multiple, opposing, intertwined, and intersected discourses that can initiate violence, racism, bigotry, and sexism. Such discourses convey their ideas very tactfully, subtly, and implicitly using different techniques, such as various linguistic features or partial representation of some facts. In Iran, politicians might use unsourced evidentials for election-related purposes; this might happen in other parts of the world for various purposes (e.g., recent demonstrations against COVID-19 vaccines) beyond elections. Janks (2010) warns that in South Africa the return of discourses of Apartheid is impending and educators should be watchful of it. One way that discourses establish their sovereignty is through texts and their linguistic features. Similar to discourses of Apartheid, other discourses such as racism, sexism, or patriotism are socially constructed, vibrant (Wong et al., 2021), and reinforced through texts and linguistic features in them (Zaini, 2022a). Accordingly, making our students familiar with the workings of unsourced evidentiality, regardless of geographical location, can have long-term and deep effects on our societies and for the sake of protecting them.

Finally, group discussions can have constructive impacts on reading against unsourced evidentials. While our participants did not extensively exchange their ideas on the use of unsourced evidentiality in texts, they had discussions on other features, such as appropriate adjectives, nouns, and adverbs, with productive effects on reading texts critically. Moreover, examples 10 and 11 above indicate that students’ cognizance of unsourced evidentiality and group discussions can be constructive. Hence, teachers’ assistance in raising students’ cognizance of the existence and workings of unsourced evidentialities, as well as group discussions, can be helpful. Unsourced evidentials might appear in different texts or languages (Huang & Wang, 2021), and students should be prepared to consume them critically.

Conclusions

The current study has revealed that unsourced evidentials can influence participants’ critical readings of texts. In effect, due to being exposed to the discourses in Iranian society, predominantly relying on oral culture, our participants did not resist the use of unsourced evidentials in texts. It is suggested that explicit education on the use and workings of unsourced evidentials is needed to empower students to improve their RC skills; explicit education can yield fruitful results as it is associated with improving students’ cognition (Zaini & Mazdayasna, 2015). Moreover, the use of marker ‘−i’ ostensibly reinforces the dominance of unsourced evidentials in the discourse of the Persian language. The participants showed a tendency to use unsourced evidentials to establish an argument.

This study has some limitations as well. First, due to the exploratory nature of this study, we had to conduct a case study. Hence, there is no claim to generalization. However, future studies can involve a larger sample of participants or study its workings in other languages. Second, our participants were in their early phase of their Ph.D. studies or they were Master’s students. We highlight that these students have been in Australia for a limited time. Hence, the western education they have been receiving in Australia might have had a limited impact on their attitudes, readings, and cognizance of unsourced evidentiality. We are not sure whether and how late-phase Ph.D. candidates might deal with the use of unsourced evidentials in texts, or, as mentioned above, how life after a Ph.D. with non-academic careers will respond to this type of inquiry. We would further suggest some future studies that focus on similar investigations, particularly in non-democratic countries to evaluate how students’ interpretations of texts is influenced by the use of unsourced evidential.

Acknowledgments

We acknowledge the active participation of our anonymous participants whose contribution yielded rich and constructive data for our whole project. We also highly appreciate the insightful comments of two anonymous reviewers whose comments improved the quality of this journal article to a great extent. Finally, we are grateful to the editors of Critical Inquiry in Language Studies for considering our manuscript.

Disclosure statement

No potential conflict of interest was reported by the authors.