Language assessment in EMI: unravelling the implicit-explicit dichotomy

: English Medium Instruction (EMI) is a content-focused approach in which language learning seems to have been sidelined. However, there appear to be some inconsistencies between EMI language objectives and actual practices that are re-ﬂ ected in the assessment phase. The literature has revealed that an implicit-explicit dichotomy is being used to explain the role of language in EMI; here explicit refers to the declared language outcomes as opposed to implicit , meaning implied but not stated. In this study, the description of a Linguistics course taught through the medium of English, the ﬁ nal assignment and its evaluative rubric are critically examined. The overall aim is to discuss discrepancies between language objectives and language assessment and identify atheoretical framework that could explain the implicit-explicit dichotomy in EMI. The present study raises the possibility that the discrepancies between language objectives and assessment criteria could be described in terms of teacher practices and beliefs that diverge from the declared language policy. Findings further indicate that the implicit-explicit dichotomy depends on academic literacies constructs, which have implications for the students ’ development of linguistic tacit knowledge, which is learnt experientially but cannot be easily codi ﬁ ed.


Introduction
EMI established itself as a phenomenon of globalisation back in the 1990s (Dearden 2015;Wachter and Maiworm 2008) when universities across Europe started to offer programmes taught in English in order to meet the ever more competitive forces of global markets (Breeze and Sancho Guinda 2022).Around the globe, English was accepted as the increasingly dominant language in many sectors and EMI was seen as an internationalisation drive.In fact, EMI programmes were initially implemented with the aim of facilitating cultural exchanges and mobility among students and staff at universities; afterwards, the main reason for the development of EMI programmes was to improve career opportunities for local students (Wachter and Maiworm 2008).
When considering definitions of EMI, content is often spotlighted.For instance, in his well-known definition, Macaro (2018) describes EMI as "[the] use of English language to teach academic subjects (rather than English itself) in countries or jurisdictions where the first language (L1) of the majority of the population is not English" (p.12).This definition emphasises content learning through the medium of the English language; however, it leaves the question of language learning open.Macaro (2018) supports the idea of English learning as being incidental; he observes how greater exposure to the English language through EMI programmes can only mean improvement in language proficiency, despite not focusing explicitly on language learning and teaching.
However, previous research (e.g., Block and Mancho-Bares 2021;Dearden 2015;Macaro 2018;Rodriguez-Melchor and Walsh 2022) demonstrates that there is no clear consensus on how content-focused as opposed to language-focused EMI programmes are.While some definitions of EMI (Chapple 2015;Rose et al. 2019) do highlight the double benefit of learning both language and content in EMI programmes, English learning is confined to an implicit outcome.Some definitions do not even make explicit reference to language outcomes (Hellekjaer 2010); others, taking into consideration the language objectives of EMI, highlight the fact that language learning is deprioritised.Regardless of the nuances, four aspects regarding the language in EMI provision seem to be constant (Pecorari 2020): 1. English is the language used for communication; 2. English is not the subject of instruction; 3. Language learning is not one of the intended outcomes; 4. English is not the mother tongue for many of the participants in an EMI setting.
Similarly, when comparing EMI to other language and content teaching approaches, it seems that EMI's focus is explicitly on content.In their continuum of approaches, Galloway and Rose (2021), place EMI at one of the ends, as the most content-focused.EMI is conceived here as the least language-focused approach, in which language learning goals are not considered.Nevertheless, the authors (Galloway and Rose 2021) acknowledge the fact that English language learning is often one of the reasons why EMI programmes are implemented and why students enrol in them (2021: 34).The literature seems to confirm this view as results from numerous studies (Dang et al. 2022;Reynolds et al. 2022 among others) have reported significant gains in vocabulary acquisition in students learning a subject through EMI.Despite the emphasis on vocabulary acquisition, the extent to which EMI has a positive effect on language learning is still a debatable issue (Graham et al. 2018).
There appears to be consensus, though, on the incidental nature of language learning, which is believed to take place implicitly by simply exposing students to English (Weinberg and Symon 2017).In other words, English is not taught but it is expected to be learned (Pecorari and Malmstrom 2018).This view belongs to an emergent trend, which sees language learning as an implicit outcome of EMI: language aims are not explicitly stated in EMI programmes but language learning is implicitly acknowledged as a result of English medium provision (Pecorari and Malmstrom 2018;Weinberg and Symon 2017).It follows that the phase of assessment, in which the students have to use the English language to demonstrate their knowledge, may create complexities.The controversy essentially entails the mismatch between stated objectives, which arguably do not comprise language learning outcomes, and assessment criteria, which potentially evaluate competencies in the English language.
The purpose of this paper is to identify possible discrepancies between the declared language objectives as stated in the course description and language assessment criteria as listed in the evaluative rubric; the overall aim is to discuss a viable theoretical framework that could explain the mismatch between explicit and implicit language objectives and assessment criteria.Thus, the questions that motivated this study are: RQ1: What are the discrepancies between the course objectives and the assignment requirements regarding language learning?RQ2: What are the discrepancies between the evaluative rubric and the assignment requirements regarding language learning?
Drawing on the existing literature on assessment in EMI and more broadly in higher education contexts, the first section of the paper offers an overview of the tension between language and content in EMI assessment, which underlines their interwoven nature.The second section of the paper describes the type of essay required in a Linguistics course provided through EMI at a university in northern Italy as well as its final assignment.This section aims to bring to the fore the role of language in EMI assessment, which often remains implicit in the course description, as opposed to the explicitness of language assessment criteria.The following section focuses on the rubric used for evaluation and analyses its linguistic components against the relevant literature.The aim is to identify inconsistencies between the definition of the language objectives, the requirements for the assignment and the rubric, paving the way for a clarification of the implicit-explicit dichotomy in EMI.
The last section aims to discuss this dichotomy against Spolsky's (2004Spolsky's ( , 2012) ) language policy framework, Borg's (2003Borg's ( , 2006) ) language teacher cognition constructs and Lea and Street's (2006) academic literacies model. 1 2 Assessment phase in EMI Language policies in EMI environments have been found inadequate to support students learning and offer guidance to teachers and tutors (Heron et al. 2021;Lasagabaster 2018); this has led EMI practitioners to develop practices based on their beliefs about English learning and teaching, which implicitly reflect their own language policy (Heron et al. 2021;Yuan et al. 2020).As Heron et al. (2021) point out, "These language practices become the implicit, unstated language policy of the individual teacher, which may potentially become a university-wide unstated policy" (2021: 3).Considering these shortcomings in language management, teachers' beliefs and attitudes become especially relevant in the assessment phase.
Some studies (Airey 2012;Ball and Lindsay 2013;Block and Moncada-Comas 2019;Doiz et al. 2019) have disclosed that most EMI teachers do not consider themselves language teachers.They perceive themselves as incapable of assessing and evaluating students' language production, despite identifying English proficiency as a positive outcome for both lecturers and students (Aguilar 2017: 732).The marginal role that language teaching appears to have within EMI reflects the wider debate on the status of English for Academic Purposes (EAP) in higher education settings.In EAP language is seen as a way to achieve discursive academic purposes (Bruce 2019) and is consequently perceived to be a service to support students' entry into the academic community (Ding and Bruce 2017).In a similar vein, EMI teachers appear to identify themselves as content teachers (Aguilar 2017) and say that language learning is mainly incidental and remains unassessed.Nevertheless, according to a study by Karakas (2017) students in EMI programmes seem to appreciate their teachers' role as language educators.Some of the participants in Karakas' study (2017) claimed that their EMI tutors took up the role of language teachers supporting their performance in English; however, this study failed to consider the teachers' perspective.
1 A academic literacy is considered the mastery of higher thinking skills, often associated with content and language learning (Shanahan and Shanahan 2008).This process is not merely individual or cognitive but is influenced by the socio-cultural context in which academic socialisation occurs (Wenger 1998).In fact, the academic literacies model (Lea and Street 2006) views the acquisition of academic literacy as a socialisation practice.In this model, becoming academic literate means sharing cultural and communicative values, which are embodied in the genres and discourses typical of each discipline.
Compared with teachers' perceived inability to act as language teachers as evidenced in other studies, Karakas' study brings to the fore an interesting aspect of EMI provision, which is the possible mismatch between EMI teachers' declared and enacted practice.
In fact, there appears to be some misalignment between what EMI teachers claim and their actual practice as it is experienced by the students.This is especially evident in language assessment, which occurs in EMI despite teachers claiming otherwise.Recent studies (Block and Mancho-Barés 2021;Mancho-Bares et al. 2022;Rodriguez Melchor and Walsh 2022;Sahan and Sahan 2022) have found that both in oral and written examinations, content lecturers are always assessing the language, focusing on phonology, morpho-syntax, fluency, correctness, coherence and register.Language assessment seems to occur unavoidably: teachers are "always somehow assessing their students' use of language […] starting from the basic assumption that if a text is not comprehensible, the student who wrote it simply could not pass the exam" (Rodriguez Melchor and Walsh 2022: 364).This covert practice is evidenced also in studies on corrective feedback in EMI (Mancho-Bares and Aguilar-Perez 2020), in which teachers refuse any responsibility for the teaching of English, but in the assessment phase correct students' linguistic mistakes, especially lexical ones, offering feedback on linguistic issues.Hence, some authors speak about implicit and explicit assessment, without properly clarifying its meaning.
These inconsistencies in assessment criteria have been amply discussed in the general literature on assessment in higher education (Bloxham et al. 2016).Findings have revealed that assessors use additional criteria to those originally stated and draw on them, whether consciously or unconsciously.Implicit assessment criteria have been identified as those that are not stated in official documents or shared with the students or other assessors (Hunter and Docherty 2011;Shay 2005).In the context of EMI, the dichotomy between explicit and implicit assessment seems to be solely related to the language outcomes of a course: despite not being explicitly stated as an outcome, English language use is nevertheless expected and consequently assessed.In other words, students' performance in EMI courses is evaluated while also taking into consideration their language competence.The assessment phase appears to surface teachers' beliefs concerning language proficiency as well as institutional expectations.
A way to disclose EMI teachers' approach to language learning is by analysing the evaluative rubrics.Rubrics have been extensively implemented in higher education with the aim to improve transparency in marking, secure appropriate standards and consistency in grading (Bloxham et al. 2016).The literature (Block and Mancho-Barés 2021;Rodriguez Melchor and Walsh 2022) has confirmed what was previously discussed: language learning resurfaces in the assessment phase.In Block and Mancho-Barés' study (2021), all items in the rubrics used for EMI courses are related to language and communication despite the teachers stating that they are only teaching content.Similarly, in Rodriguez Melchor and Walsh's study (2022), the teachers claimed that they did not assess according to language criteria.However, once asked to attribute value to the importance of certain language criteria in rubrics, they assigned high scores to almost all linguistic parameters.This attitude seems to confirm the contradictions in EMI assessment between what is explicitly stated in the course objectives and what inevitably occurs, i.e. language assessment.
The tension between explicit and implicit objectives might be due to EMI's alleged nature.EMI has been conceived as a content-focused approach in which language teaching and learning are sidelined (Galloway and Rose 2021).Institutions have adjusted to this view in order to improve internationalisation and staff mobility.The literature to date, however, appears to question the initial assumption according to which EMI stands at one of the extremes in Galloway and Rose's continuum (Figure 1).As discussed, teachers' beliefs highlight the relevance that language assumes in EMI.Revisiting the position that the English language has in EMI could mean redefining EMI as ILCHE (Integrating Language and Content in Higher Education 2 ), so to shift the focus on both content and language.2 Galloway and Rose's (2021) spectrum of approaches to content and language teaching presents on one extreme English language Teaching (ELT), in which the focus of instruction is on language, on the other extreme EMI, implying that its focus is on content learning and teaching.ILCHE is conceived as the parallel approach to Content and Language Integrated Learning (CLIL) in higher education.CLIL and ILCHE entail that the focus of instruction is on both content and language.

Essay writing in an EMI linguistics course
This section of the paper aims to describe the type of assessment required in a Linguistics course taught at the Master's level at a university in northern Italy.The Linguistics course is blended, offering on-campus and online classes, for a total of 30 h of lessons.Upon successful completion of the course, the students receive six ECTS (European Credit Transfer System) points towards their Master's degree.The course is entirely taught through the medium of English; the assessment is equally done in English.This course belongs to a wealth of programmes taught through EMI offered at the aforementioned university.As stated on the university website, international programmes are devised as an effort to make the academic community more international and more internationally minded, thus meeting the dual aim of EMI provision: attracting international students and developing English learning among the local community.The university provides a webpage on which international programmes are described and links on how to access them are listed.However, confirming Marcos-García and Pavón's (2018) study results, the language policy documents seem to be unavailable on the web; thus, this paper cannot provide a detailed account of the aims of EMI provision in this context.
Both Italian and international students attending this course are required to have a certified B2 level of English according to the CEFR (Common European Framework for Languages, Council of Europe 2001).Despite general English and EAP modules not being mandatory for students, they are offered throughout the academic year as part of the curriculum of studies representing significant language learning opportunities.In terms of academic literacy in English, it is worth noting that some competencies could be considered acquired by postgraduate students; in fact, entry requirements for the Master's degree comprise an undergraduate degree in the field of Linguistics or Foreign Languages and 12 ECTS points in the English Language.In line with these high expectations for students, teachers in this university are required to certify their C1 level of English to be eligible for EMI teaching.This specific context, in which both teachers and students are supposed to have a good command of the English language, might be influential in curriculum planning, setting high standards in terms of linguistic performance for both teachers and students.
The main aim of the course is to acquire knowledge in the field of Educational Linguistics, especially concerning the advancements in the discipline and the role of multilingual practices.As regards the outcomes, the programme of the course expects students to gain insights into the historical developments of language education, both at the national and the international level, and to be able to plan learning paths considering different learning approaches and contexts.It is also Language assessment in EMI expected that students attending the course will be able to make informed decisions regarding their learning paths based on relevant knowledge.A final section of the course programme explicitly states that students should be able to express their ideas clearly and unambiguously (it is not specified whether in writing or orally).
The assessment of the course comprises two parts, both weighing 50 % of the final grade: an essay and a written exam.The essay aims to assess the ability to analyse and discuss one of the current theories in the field of educational linguistics, and a written exam made up of three questions checks knowledge and awareness of the most current trends in this field.This paper will look at the essay component in the assessment.
The assessment guidelines for the essay consist of three parts: the first part is a description of the assignment followed by a reading list; the second part presents some guidelines on how to write the paper; the last part outlines the format in which students should write their paper.
According to the assignment description, the essay requires the students to analyse and discuss one of the main trends in language education and multilingualism.Students are required to choose an article or a book chapter from a list given by the teacher and another one of their choice: the latter should confirm, contradict or explore further the argument of the former text.The writing guidelines, besides stating the parts which the essay should comprise (title, introduction, theoretical background and argumentation, analysis of main points, conclusion, final comment, and bibliography) also indicate the contents expected in each section.In brief, the assignment is an argumentative essay in which students are expected to develop an argument discussing two sources from the relevant literature.
The final section of the writing guidelines presents some language features that the essay is supposed to follow.It is worth noticing that the language objectives are mentioned only at the assessment stage of the course.While the course programme claims that one of the course prerequisites is that students should be able to express themselves clearly, no more indication is given as regards specific language skills.In contrast, the writing guidelines explicitly require that students use a formal register (avoiding colloquialisms, such as the first person, contractions, informal words and expressions), use punctuation effectively, focus on coherence and cohesion in the argumentation, and pay attention to clarity.
Overall, it seems that the main course aims are to make the students interact with the more recent trends of educational linguistics, developing both knowledge and skills in teaching planning.Language aspects appear to be overlooked or taken for granted in the course programme, whereas, the assessment stage of the course explicitly expects that the students use English at a proficient level, showing a strong command of academic writing skills.It remains uncertain whether this level of English competence is a prerequisite or an objective of the course, and what weight language has on the essay evaluation.
4 The discrepancies between the course objectives and the assignment requirements regarding language learning Bloxham and West (2007) describe written assignments, and more specifically examinations, as complex tasks, which involve the students' capability of adapting to new social practices and using new forms of expression in order to successfully become part of an academic community.This view on writing draws on the academic literacies approach (Street and Lea 2000), which sees literacies as social practices, requiring the students to deploy a repertoire of linguistic practices appropriate to each setting.It is evident that the essay described here reinforces this view of an academic community of practices because it requires the students to engage in a specific academic form of discourse.
As discussed above, the students are expected to use and master a specific essay structure described in the guidelines.It is unclear though whether and to what extent the students are familiar with this structure; this uncertainty reiterates the tension between explicit and implicit outcomes in EMI courses, which is also present in the assessment phase.The course objectives explicitly state that students are expected to express their knowledge clearly, but it is debatable whether the objectives also comprise the ability to write academically in English.Nevertheless, it can be argued that the entry level for an EMI course implies a sufficient command of the English language to perform this type of task.A recent study on entry levels for EMI courses in Italy (Cicillini 2021) has found that despite not adopting a homogeneous strategy for certifying English language proficiency, most Italian universities require a B2 (Council of Europe 2001) certificate in English for entry at Master's level.Entry levels in English have been set to ensure high educational quality in EMI programmes and avoid unbalanced classrooms in terms of linguistic ability (Lasagabaster 2015).This might also imply that students are expected to master a certain level of academic English, especially because in Italy they are supposed to reach the B2 level by the end of their school studies.
The discrepancies between course objectives and assessment requirements might be informed by the notion of tacit knowledge.In the assessment described here, tacit knowledge refers to the linguistic requirements of the essay, which entail both the academic language and the structure of the essay.O' Donovan et al. (2004), referring to what is expected in assessments, define tacit knowledge as "that which is learnt experientially or in terms of its incommunicabilityknowledge that cannot be easily articulated" (p.328).They also suggest that teachers use tacit knowledge in their marking.In fact, the ability to write satisfactorily within the academic community entails learning tacit knowledge, for example, what is expected in assessment.Moreover, the Linguistics course described here belongs to the curriculum of Language Sciences, which presupposes that the students use English throughout their academic career (i.e.sitting exams and writing their final thesis entirely in English) and are familiar, though not properly trained, with academic writing in English.
It is evident that what students are expected to learn tacitly, rather than explicitly, encompasses also the broader idea of academic literacies: writing a satisfactory assignment involves being socialised in new practices and forms of expression as well as using appropriate study skills (Lea and Street 2006).Becoming socialised in an academic community is essentially measured through the capacity to write adequate assignments and examinations.This relationship between academic writing and literacies reflects the fil rouge between language and academic socialisation.Furthermore, these findings suggest that on a practical level, the definition of objectives in an EMI course should also consider those linguistic prerequisites that are expected from the students.A clear understanding of the linguistic expectations could promote students' participation in the non-mandatory general English and EAP modules, which aim to improve language proficiency and enable students to write academically in English.
In sum, it appears that the essay represents a form of literacy that the students are expected to acquire, implicitly or tacitly, belonging to a specific academic community.Although not explicitly stated in the course objectives, this academic literacy surfaces in the assessment phase.The assignment requirements explicitly foreground the language dimension, thus confirming the role of linguistic tacit knowledge in the EMI course.The pivotal role of language, and more specifically of academic English, seems to shift the focus of EMI instruction towards both content and language learning.Despite the course description not comprehending language learning objectives or language teaching aims, language learning is expected as the assignment requirements demonstrate.This double focus on both language and content poses some challenges in the comprehensive definition of EMI, as it is conceived according to Galloway and Rose's (2021) continuum; it could be argued that EMI is a context-sensitive approach, which is unlikely to fit in an overarching term.

The rubric
Essay evaluation in the EMI course described here is carried out using a rubric.The rubric is written by the lecturer and is mainly used by the tutors responsible for the marking of the essays; it is not, however, shared with the students.The performance rating goes from fail (which corresponds to less than 18 points) to 30 points: 1. 18 to 21-grade range describes a work that is sufficiently developed but with several weaknesses; 2. 22 to 24-grade range reflects fairly good work with several aspects to improve; 3. 25 to 26-grade range means that the essay meets all the assessment requirements with some room for improvement; 4. 27 to 28-grade range reflects a very good response; 5. 29 to 30-grade range describes an excellent piece of work.
The rubric is composed of five dimensions: 1 relevance of sources, 2 understanding and interpretation of sources 3 logical presentation and structure, 4 unity, cohesion and clarity of writing, 5 evidence in justifying claims.Performance descriptors are not included in the rubric, nor made available elsewhere.However, the lecturer provided some guidelines for the tutors responsible for the marking.The guidelines refer exclusively to the meaning of each criterion and aim to guide the tutors in evaluating the essays: 1. Relevance of sources: the students should make reference to only two papers as requested in the assignment.The second paper, which the students choose autonomously, should have academic credibility.2. Understanding and interpretation of the sources: the essay should demonstrate a thorough understanding of the two articles; the articles should be mentioned appropriately using the APA system of citation.Only the two papers should be cited in the reference section of the essay (if students mention other authors, which are well-known sources in the field of educational linguistics, they should not write them in the reference section).3. Logical presentation and structure: the analysis of the articles and the argumentation are clear and fluent; the paper's structure meets the assignment's requirements.4. Unity, cohesion and clarity of writing: language mistakes should not have a negative impact on the evaluation unless they hinder understanding.The lack of linking words and poor cohesion and coherence should be evaluated negatively.
The essay should present a formal academic style.The first-person singular is accepted exclusively in the final comment.
5. Evidence in justifying claims: in the analysis section of the essay claims should be justified on the basis of the two sources chosen; personal comments are accepted only in the final section (personal comment).
Evidently, students are expected to be able to have a strong command of academic English, both in terms of lexis and syntax (analysis and argumentation should be clear and fluent) and in terms of style and genre (cohesion and coherence are evaluated as well as the use of a formal register); students are also expected to justify their claims based on scientific sources.However, the rubric guidelines explicitly state that language mistakes should not have a negative impact on the evaluation unless they compromise understanding.It is unclear what language mistakes are not supposed to be evaluated and how they should be considered in terms of the final mark.Furthermore, it could be argued that the usage of linking words as such does not guarantee coherence and cohesion in a text.
6 The discrepancies between the evaluative rubric and the assignment requirements regarding language learning In this study, the term "rubric" refers to a scoring tool, which presents the specific components of an assignment and the description of acceptable and unacceptable levels of performance (Stevens and Levi 2005).As described by Reddy and Andrade (2010), a rubric is usually in the shape of a grid: the list of components to be assessed (known as "dimensions") is usually placed in the leftmost column; a column is set up for every performance level to be assessed; the scoring strategy normally appears as a numerical scale in the top row.The description of the levels of performance also appears in the grid.However, rubrics do not always follow this design and teachers create other types according to necessity.Flexibility and adaptability make the rubric a context-specific tool, which embodies attitudes and beliefs about assessment in a specific setting.As outlined in Section 3, the EMI rubric presented here considerably differs in its design from Reddy and Andrade's (2010).This rubric is presented in a discursive form rather than in a grid and interestingly, the desirable qualities of the final assessment are highlighted but less consideration is given to the lower gradation of quality for each dimension.
The traditional literature on the usage of rubrics in higher education has foregrounded the influence that rubrics could have on teaching and learning, which has been described as a "washback effect" (Alderson and Wall 1993: 115).It appears that the evaluative criteria of the rubric trigger the definition of the course contents and its methodology, shaping the curriculum and the teaching and learning process as a whole.Rubrics have also been found to have an influence on setting learning goals as well as evaluating students' performance (Andrade 2005(Andrade , 2000)).According to Andrade (2005), the process of drafting a rubric encompasses the main goals and objectives of a course or a module, on which the assignment is supposed to be based.While a rubric is conceived as an assessment tool that sets the criteria against which an assignment might be evaluated, its scope and influence seem to go well beyond the mere assessment phase, influencing the whole curriculum.This double purpose is highlighted in the literature; some studies (Campbell 2005;Tunon and Brydges 2006) have focused on the mere evaluative purposes of using rubrics in Higher Education, while other research (Schneider 2006;Song 2006) has delved into their broader educational scope.
In the rubric discussed here, language proficiency is pivotal.Students are required to master the English language at a high level in order to fulfil the assignment requirements.This is despite the fact that there is no mention of writing academically in English in the course objectives but only of being able to express oneself clearly.The rubric and the assignment are coherent with one another; both show a significant focus on academic writing in English and on mastering English for Academic Purposes.However, when comparing the course objectives with the evaluative rubric, inconsistencies appear: the course description focuses on the content dimension of the course, mentioning only the ability to express oneself with clarity both orally and in writing.Similarly, the criteria listed in the rubric only marginally appear to confirm the explicit objectives of the course (e.g., in the ability of the student to choose the most relevant sources and interpret them).Language objectives are, in short, implicit in the course description: they are not stated initially but unavoidably resurface in the assessment and evaluation phase.It seems evident that the rubric surfaces aspects related to the hidden curriculum, which do not appear in the module description.It remains to be understood whether proficiency in academic English is implied due to the expected language competence of the students, or whether the course description has to follow guidelines at the institutional level, which do not comprise language outcomes in EMI.
As confirmed in the literature on EMI assessment (Rodriguez Melchor and Walsh 2022), implicit biases have been found in relation to linguistic criteria.The teachers interviewed by Rodriguez Melchor and Walsh asserted that there are linguistic implicit biases related to the readability of students' writing; despite stating that they do not assess according to linguistic criteria, linguistic parameters are seen as pivotal to the understanding of the text.In a similar way, the discrepancies highlighted in the Linguistics course described here might be linked to implicit assumptions regarding students' ability to write and express themselves in English.As noted previously, teachers' beliefs on language learning in EMI might have compensated for the lack of language policy (Heron et al. 2021;Yuan et al. 2020) and become evident in the assessment phase.In contrast, the official description of the course seems to adhere to the institutional and general requirements of courses taught through the medium of English (i.e.content-focused, language learning only incidental).These findings provide some tentative evidence of the institutional pressure to implement EMI courses at universities without considering the necessary linguistic support.Therefore, there are calls for a constructive dialogue among the various stakeholders, involved in the planning and delivery of EMI courses, in order to take into consideration students' linguistic needs.

Implicitness in EMI
The evidence behind the implicit-explicit dichotomy appears to draw mainly on three interrelated constructs: language policy (Spolsky 2004(Spolsky , 2012)), teacher cognition (Borg 2003(Borg , 2006)), and the academic literacies model (Street and Lea 2000).Figure 2 illustrates an integrated framework, which includes multiple components that intersect, inform and influence each other.
As previously discussed, the literature has pointed out that academic institutions adopting EMI programmes have not sufficiently invested in language policy (intended here as language management).This lack of institutional support might have led to teachers relying on their beliefs about language learning and adapting their practices accordingly.In Spolsky's model (2004Spolsky's model ( , 2012) ) beliefs and values about the language are identified as one of the components of the language policy itself.They are related to language practices and might affect or be affected by language management; in other words, language management as well as language practices can modify or be modified by language beliefs.As Spolsky (2004) observes, language management, or the guidelines and laws that regulate a certain speech community, might contradict language beliefs and practices.In the present study, a certain level of tension is reflected in the discrepancies between the description of the Educational Linguistics course and the assessment phase where the teacher's values regarding language appear to come to the fore.
Along the same lines, Borg's teacher cognition (2003,2006) refers to those unobservable constructs that teachers develop as a result of the environment in which they operate and become teachers' beliefs.They may or may not be reflected in the teaching practice: mismatches have been found between what teachers claim to believe and what they do in the classroom.These incongruities seem to be caused by the influence of school policies and curriculum mandates on both cognition and practice (Phipps and Borg 2009;Roothooft 2014).As Borg (2003) points out, the influence of contextual factors might be evident in the discrepancies between cognition and practice and more specifically in the teachers' inability to adopt practices that reflect their beliefs.In this study, similar incongruities are evident in the disproportion between the language focus in the assessment phase and the course objectives.
Furthermore, the academic literacies model highlights also the implicit/explicit dichotomy in EMI as it foregrounds the "institutional nature of what counts as knowledge in any particular academic context […] involving both epistemological issues and social processes" (Lea and Street 2006: 368-9).This model views students' writing as an issue of epistemology and identity rather than merely skills acquisition or academic socialisation and considers the broader institutional discourses and genres associated with academic writing (Lea and Street 2006).Student writing can be understood not only by taking into consideration institutional practices and identities but also power relations, intended as societal requirements, curriculum mandates and availability of resources (Borg 2003).EMI programmes, in fact, stem from the growing momentum towards internationalisation in academia (Schmidt-Unterberger 2018), which shapes curricula according to higher globalisation stakes.In this context, the English language assumes a vehicular function, that is, it is viewed as the language of instruction, rather than a subject itself.Language learning objectives are not explicitly stated but implicitly acknowledged by the academic community (in other words, there are linguistic implicit biases in EMI).Students become acculturated into the discourses and genres of the discipline and are expected to master them alongside the language.

Conclusions
This study set out to investigate the implicit-explicit dichotomy in the assessment phase of an EMI course in Linguistics.In this investigation, the aim was to discuss the evidence behind the choice of an essay for assessment and the accompanying evaluative rubric.The most recent literature appears to be using the implicit-explicit dichotomy without properly defining the meaning and the scope of the two terms.This issue has grown in importance in light of recent findings on language learning (Pecorari and Malmstrom 2018;Weinberg and Symon 2017) and assessment (Block and Mancho-Barés 2021;Rordiguez Melchor and Walsh 2022;Sahan and Sahan 2022) in EMI.
In EMI explicit appears to mean clearly stated and acknowledged.In the example described in this paper, language learning objectives are barely stated in the assessment nor are they in the course description, but they are strongly affirmed in the rubric used for the assignment evaluation.In contrast, implicit seems to refer to objectives, which are not stated but generally expected.According to the literature, not only is language learning implicit but so are also language objectives.As previously discussed, this dichotomy creates a misalignment between EMI lecturers' assumptions and practices: language learning objectives appear explicitly in the evaluation rubric, contradicting what was previously stated.
In line with the literature, the study has found that language objectives are implicit in the course description but they resurface in the assessment phase.An implication of this is the possibility that tacit knowledge regarding language learning is expected but not expressed in the description of the course.This study has raised questions about the nature of implicitness in EMI: the language policy constructs (which in this proposed framework encompass also language teacher cognition) alongside the academic literacies model might explain this issue in terms of identities and academic acculturation.Considering EMI's pull for internationalisation, power relations could also held accountable for the shaping of the curricula.Considerably more work will need to be done to determine the epistemology of the implicit-explicit dichotomy in EMI, especially as regards other forms of assessment.

Figure 2 :
Figure 2: Proposed conceptual framework for explaining the implicit-explicit dichotomy in EMI.