Evaluative language in undergraduate academic writing: expressions of ATTITUDE as sources of text e ﬀ ectiveness in English as a Foreign Language

: The purpose of this paper is to explore whether the use of attitudinal language stands as a potential source of e ﬀ ectiveness in undergraduate academic writing in English as a Foreign Language (EFL). In order to achieve this purpose, interpersonal features of a corpus of essays written by Mexican undergraduate students of English Language and Literature were analyzed. The model of APPRAISAL (Martin, James R. & Peter R. R. White. 2005. The language of evaluation: Appraisal systems in English . Basingtoke: Palgrave Macmillan) was used to trace and contrast ATTITUDE resources of a ﬀ ect,judgement and appreciation in academic essays in relation to the grades they were granted by university professors at di ﬀ erent levels of instruction. The results of the study con ﬁ rm a signi ﬁ cant relation between the use of resources of ATTITUDE and the perceived (in)e ﬀ ectiveness of the analyzed texts, as well as factors which potentially determine the nature of such relation and pose relevant implications for academic writing instruction in EFL in the context of the analyzed corpus.


Introduction
Writing academically in English as a Foreign Language (EFL) is widely recognized as central to undergraduate education due to its role as the main means of production and dissemination of knowledge in international contexts (Nesi and Gardner 2012).
Academic writing is also considered as an activity which involves complex sets of skills that even expert writers keep on improving after continued experience (Chazal 2014).As part of such complexity, the appropriate expression of interpersonal meanings has been found to be one of the most challenging components when learning how to write academically (Hood 2010), mainly because of the pedagogical challenges involved in an appropriate induction of learners into the linguistic conventions developed by specialized discourse communities to produce interpersonal meanings.
Different from skills and competencies related to the communication of factual information and conventional paradigms of textual organization, construing interpersonal meanings involves the use of linguistic resources to express interaction among speakers, as well as the communication of the feelings and points of view they intend to share (Moss 2011).Interpersonal meanings also convey social roles and evaluations (Hyland 2005), as well as the speakers' expression of a subjective reality, which conveys social relations, personal values, and different kinds of opinions they want to share with other speakers (Perales et al. 2012).Additionally, the expression of interpersonal meanings relates to authorial presence, also known as stance, in academic texts; "how writers present themselves and express their own views and judgements" in relation to others (Candarli et al. 2015: 193).Because of their discourse semantic nature, the workings of these dynamics and their rhetorical effects might make it harder for learners of EFL academic writing to acquire, observe, and evaluate than grammatical and organizational features, as attested by observations like Lancaster's (2014), who has referred to interpersonal dynamics as recurring patterns of language use that are difficult to notice from casual scanning.
As explained by Ken Hyland, research has shown that fundamental challenges in teaching and learning interpersonal conventions in writing are found in two main sources of difficulties.For one thing, despite its actual social nature, academic writing may still be regarded as objective, rational and impersonal, which results in a neglection of the fact that academic, disciplinary practices involve variations in systems "of appropriate social engagement with one's material and one's colleagues" (Hyland 2004: 11).Additionally, cultural factors that shape students' expectations, strategies, beliefs and, in general, schemas of knowledge and the whole process of learning also present learners with important challenges; besides the acquisition of grammatical structures, lexical resources, and notions of textual organization, learners writing in EFL have to deal with differences between diverse cultural conceptualizations of knowledge construction and communication.While certain cultures tend to favor "an analytical, questioning, and evaluative stance to knowledge", others "have a very different perspective that favors conserving and reproducing existing knowledge" (Hyland 2003: 38).
Despite the abundant production of research this subject has seen around the globe, the exploration of the relationship between interpersonal resources and effectiveness in academic EFL writing has not received much attention in Mexico and Latin America, where most related research has focused on academic writing in Spanish L1 (Castro 2013;Castro and Sánchez 2013;González 2011;Ignatieva 2021;Navarro 2014;Valerdi 2021;Zamudio 2016).It is also worth noticing that, as it will be shown in further sections of this paper, most previous works beyond Latin American contexts have explored the interpersonal discourse of EFL writers from disciplines which require effective communication in English, but not the professional mastery or depth that is expected from future professionals of the English Language and Literature at different levels of linguistic and disciplinary instruction.Such potential mastery and depth in knowledge represent the need to analyze the workings of interpersonal language considering as many realizational variables as possible.Stemming from this, the objective of this paper is to use qualitative and quantitative attitude in academic EFL writing research methods to analyze the use of interpersonal resources in a corpus of academic essays written in EFL by Mexican undergraduate students of English Language and Literature at a public university in Central Mexico.This exploration is developed within the framework of Martin and White's (2005) model of APPRAISAL with specific focus on the system of ATTITUDE and the instantiation variables of attitudinal category, realization, explicitness, trigger, and authorship.In this work, I compare the features of attitudinal realizations in low-, middle-, and high-graded essays in order to understand the relationship between the use of attitudinal language and the perceived effectiveness of the essays where it is instantiated.Additionally, the features of the evaluative language in the corpus are observed in relation to the instructional and linguistic experience of the undergraduate authors of the texts, which were classified as basic, intermediate, and advanced.
In the following section, a brief account of the system of ATTITUDE is presented with particular focus on the categories that have been analyzed in this research.This is followed by an account of previous works that have explored the relationship between attitudinal language and the (in)effectiveness of academic writing by undergraduate EFL learners from different linguistic and cultural backgrounds.Then, a description of the mixed methodological approach which was applied in this research is presented.Finally, I report the most significant findings of this work in order to discuss their implications to teaching academic writing in EFL in the Mexican context.
2 The APPRAISAL model and the system of ATTITUDE Martin and White (2005) developed the APPRAISAL model as an extension of the analysis of the interpersonal metafunction proposed in Halliday's Systemic Functional Grammar (1985, 2004).While Halliday describes this metafunction at the lexicogrammar stratum in terms of the clause as an exchange through the system of MOOD, the APPRAISAL model approaches the expression of feelings, points of view and social relations at the stratum of discourse semantics through the systems of ATTITUDE, ENGAGEMENT, and GRADUATION.The first one involves the discursive negotiation of emotion-related meanings (feelings, moral/ethical judgements, and aesthetic responses), while the second one relates to the speaker's positioning in relation to other voices.The system of GRADUATION has to do with the resources deployed to intensify and attenuate realizations of ATTITUDE and ENGAGEMENT (Martin and White 2005).
This paper is concerned with the system of ATTITUDE, which encompasses the linguistic expression of emotions and their institutionalization as opinions related to the notions of ethics and aesthetics.The ultimate objective of using these resources is to share subjective points of view for the audience to consider them and, if the discourse is successful, adhere to them.In this sense, the expression of ATTITUDE is central for speakers' argumentation of their positions and the representations of the world they communicate since "it is possible to adhere oneself not only to a thesis, but also to ways of thinking, seeing, and feeling" (Amossy 2009: 67-68).Thus, considering Vega's definition of effectiveness (2015) as argumentation which accommodates to the audience's attitudes and beliefs in order to increase its adhesion to the speaker's proposals, the rhetorical effects of communicating one's propositions accompanied by attitudinal evaluations may have the potential to enhance the effectiveness of the produced text beyond the scope of its objective contents.
In the model of APPRAISAL, expressions of ATTITUDE are analyzed in terms of three sub-systems: affect, judgement, and appreciation.Affect relates to the linguistic realization of feelings, considered as the most basic forms of personal reaction in human linguistic development (Martin 2000;Painter 2003;Torr 1997).These expressions are usually classified into the categories of +/− happiness ["affairs of the heart" (Martin and White 2005: 49)], +/− security (related to ecosocial well-being and sensations of anxiety and confidence), +/− satisfaction (linked to the satisfaction of personal needs and goals) and +/− desire (involving the willingness or attraction towards entities and processes).Prototypically, instances of affect are lexically realized in the form of qualities (attributes and epithets), mental and behavioral processes, and modal adjuncts.Table 1 illustrates examples of affect analyzed in the corpus of this study.The two columns on the right present information about the variables of realizationthe lexico-grammatical formand categoryattitudinal sub-typeof each example.
Resources from the judgement and appreciation sub-systems are the result of the modelling and sophistication of affectivity according to social norms and aesthetic parameters produced by social conventions, hence their status as institutionalizations of affect.These institutionalizations represent moral/ethic judgements about people, their actions, and the consequences of their actions, as well as personal considerations about the aesthetic properties of things, processes, Attribute +satisfaction We would like to highlight the kind of words […] Mental process + desire All the examples presented on this and other tables and figures are instances of evaluative language analyzed in the corpus of this research, except for *, which has been adapted from Martin and White (: ).
attitude in academic EFL writing and people.Following Martin and White's (2005) description, expressions of judgement are classified into the categories of +/− normality (how special or unique someone is), +/− capacity (how able a person is), +/− tenacity (how resolute someone is), +/− veracity (related to people's honesty) and +/− propriety (linked to people's integrity).Judgement is typically realized through qualities and modal adjuncts (Table 2).Expressions of appreciation encode aesthetic evaluations of speakers in terms of three categories: +/− reaction (emotional responses provoked by things and processes), +/− composition (perceptions of the balance, regularity, and order of things), and +/− valuation (opinions about the aesthetic and practical value of things).Prototypical realizations of appreciation take the form of qualities and circumstances (Table 3).
The examples of ATTITUDE presented up to this point correspond to inscribed realizations; expressions that encode ATTITUDE overtly or explicitly.Affect, judgement, and appreciation can also be indirectly invoked.The particularity of invoked realizations of ATTITUDE is that they demand a special effort from the audience or interlocutor to identify and process the type of evaluation the speaker is doing.Another peculiarity of this type of realization is that it involves the selection of resources pertaining to the ideational metafunction, which "gives structure to experience, and helps to determine our way of looking at things, so that it requires some intellectual effort to see them in any other way than that which our language suggests to us" (Halliday 2002: 175).Hence observations by Hood and Martin (2005) and Hood (2010) of processes and participants as efficient attitudinal invokers: ideational meanings are seldom neutral, and speakers choose non-inherently evaluative resources for their invoking potential.
ATTITUDE is prototypically invoked through lexical metaphors, processes, nominal participants, and nominalizations of processes and qualities.A notorious exception to these invoked realizations is the case of affect, which can be inscribed through mental and behavioral processes.Table 4 illustrates invoked realizations of ATTITUDE analyzed in the corpus of this work.
An additional remark on attitudinal invocation relates to potential double coding (Martin and White 2005) resulting from indirect realizations of ATTITUDE where a given evaluation from one category may be invoked by means of the inscription of another category as in (1) below.In this example, the author expresses a negative evaluation of a reviewer's knowledge (appreciation > −composition), which indirectly realizes a negative evaluation about his abilities as a literary analyst (judgement > −capacity).The possibility of finding such types of invoked realizations in discourse represents the need to make methodological decisions regarding the labelling of attitudinal realizations in an analysis, as was the case in this work.Pertinent observations are made later in the Section 4.
(1) Such comments reveal the reviewer's limited knowledge […] For the purposes of this research, additionally to the variables of realization, category and explicitness, the variable of trigger is particularly relevant.This variable indicates the stimulus that motivates an evaluation.For the nature of the essays studied in this worki.e., essays on literary works from English Literaturetriggers of ATTITUDE were related to characters, other components of the literary works commented by the authors of the corpus (atmosphere, style, plot, etc.), and entities external to the literary works about which they write in their essays (literary concepts, periods, genres, currents, other works, different authors, etc.).Table 5 illustrates instantiations of ATTITUDE with labels of their realization, category, explicitness, and trigger.

Research on attitudinal meanings in undergraduate EFL academic writing
The study of the linguistic expression of attitude and its relationship with successful EFL academic writing has received considerable attention in international contexts.Jalilifar and Hemmati (2013) analyzed a corpus of argumentative essays written by postgraduate Kurdish-speaking Iranian students of TEFL to explore whether APPRAISAL resources can be used as a reference to evaluate argumentative writing in low-and high-graded essays.These authors found that higher proportions of ATTITUDE were characteristic of more successful texts, although both groups of texts were similar in a tendency to use more judgement and appreciation than affect.Additionally, high-graded essays displayed significantly more invoked affect through nominalizations and more inscribed valuation than low-graded texts, which tended to inscribe affect through mental processes and to invoke valuation.In terms of judgement, the main contrast between both sub-corpora lied in the selection of sub-categories, with high-graded essays mainly displaying resources of capacity and low-graded texts mainly instantiating normality.Jalilifar and Hemmati (2013) conclude that, although clear tendencies showed that more successful essays tend to display more ATTITUDE, punctual instantiations suggested that "in assessing the essays as high-or low-graded, more important than the number of appraisal markers exploited in essays was how these resources were

Judgement +tenacity
Character from the analyzed literary work Invoked employed" (p.75).This conclusion coincides with observations by Hunston (2011), Hunston andSu (2019), andValerdi (2021), who have pointed at the centrality of lexico-grammatical patterns and the strategic use of evaluative language in argumentative structures over low or high proportions of APPRAISAL resources in texts.
Reaching similar conclusions, Myskow and Ono (2018) studied how Japanese undergraduate students of Law and Political Science used resources of affect, judgement and appreciation as part of their arguments' justifying evidence in a corpus of 62 biographical essays.Having rated the texts as either high or low on the basis of a self-designed rubric, these authors found that both low-and high-rated essays displayed similar proportions of general attitudinal APPRAISAL, with judgement as the most prevalent type of evaluation.Here, the relevant contrasts took place in terms of the most frequently used sub-systems in each group of texts, with high-rated essays displaying higher proportions of appreciation and judgement, and low-rated texts displaying more resources of affect.Additionally, for these researchers, what students found challenging was not the selection of particular attitudinal categories, but finding appropriate grammatical constructions for them to support their arguments; they noticed students relied heavily on attributes to inscribe conclusive evaluations about people.Myskow and Ono (2018) suggest writing teachers should not encourage learners to adopt critical perspectives through heavily polarized inscriptions, but to integrate evaluations with evidence in support of their conclusions through particular grammatical constructions.
In a work that explores the relationship between undergraduate students' lack of critical voice and limitations in the management of evaluative language, Lee (2015) compared high-and low-graded persuasive essays written by native (Australian) and non-native (South Korean, Japanese, and Taiwanese) undergraduate EAP students to identify how evaluative language contributes to academic writing success.In that study, native and non-native writers deployed judgement and appreciation ten times more frequently than affect.When focusing on appreciation resources, the author found that valuation was predominantly more used over expressions of composition and reaction, which they interpretated as a subject-specific particularity.When contrasting low-and high-graded essays, Lee found that less successful texts displayed affect-related reaction twelve times more frequently than successful essays.Additionally, she found that high-graded essays by both native and non-native writers displayed similar tendencies, with frequent use of judgement-invoking valuations.In 2008, Lee had reported very similar tendencies in a study based on a remarkably similar corpus, with high-graded essays deploying significantly more varied attitudinal resourcesmainly invoked evaluations of judgmentand depersonalized ATTITUDEin the form of nominalizationsthan low-graded texts.For Lee (2015), shaping an appropriate use of evaluative language is key in EAP instruction, where "students' exposure to the APPRAISAL system helps them to acquire the relevant English Language skills including grammar and vocabulary most effectively in a context-appropriate manner" (p.73).
In a work which supports Lee's (2015) position, Bahmani et al. (2021) propose the model of APPRAISAL as a pedagogical tool to help EAP writers show their critical stance in their texts.To support their claim, the authors developed an experimental study at an Iranian university, comparing the writing successfulness of an experimental group of 30 postgraduate students of English Language Teaching to that of a control group with the same number of students.The experimental group received writing instruction with explicit explanations and analyses of ATTITUDE and GRADUATION.On the basis of standardized pre-tests and post-tests applied to both groups, Bahmani et al. concluded that explicit instruction on the use of evaluative language enabled the experimental group to perform more successfully in contrast with the control group, which did not improve its performance significantly.
Besides research based on APPRAISAL analysis, the role of attitudes in academic writing in EFL has seen remarkable contributions developed through applications of Hyland's (2005) metadiscourse model.Despite general differences between metadiscourse and APPRAISAL, both models have been found to share remarkable similarities in their conceptualization of linguistic attitudes (Du et al. 2023): linguistic resources that "indicate the writer's affective […] attitude to propositions" via markers of surprise, agreement, frustration, importance, etc. (Hyland 2005: 53).In this paper, key metadiscourse works are worth mentioning for their special focus on contrasting less or more effective texts on the basis of their attitudinal features.
One of such works is that by Crosthwaite and Jiang (2017), who support the idea that successful attitudinal conventions can stem from instruction in academic writing.They explored how explicit instruction on the use of attitudinal metadiscourse affects the development of stance features in essays and reports.The authors investigated the development of attitude markers and other metadiscoursal resources during one semester of EAP instruction to undergraduate students from various disciplines in Hong Kong.After analyzing written samples before, during, and after explicit instruction, Crosthwaite and Jiang (2017) found that undergraduate authors had reduced their use of attitudinal markers significantly in their texts.Additionally, when they graded the same written samples, they found that interpersonal features developed through instruction were more likely to have a positive impact on evaluators.For these authors, "the ability to express a relevant and plausible stance is a crucial indicator of writing quality and development" (p.94).As challenging as this has been proved to be, such ability can be trained in undergraduate students, who enter university writing privileging their attitudes about their claims and their positions, but may be instructed to develop a new academic voice, "gaining control over the rhetorical and linguistic aspects of academic discourse via the use of a more careful, narrower, less polarising and less personal range of expressions with which to convey their attitudes on a given topic" (p.102).Lee and Deakin (2016) looked more closely at the role of attitudinal metadiscourse markers in the effectiveness of undergraduate writing in EFL by analyzing essays written by Chinese learners enrolled in their first or second writing course at a university in the United States.The essays had been rated through a standardized grading rubric either as low or high.Lee and Deakin (2016) found that authors of successful essays had a tendency to include significantly more attitudinal markers than less successful essays.Additionally, these authors observed that attitudinal resources in high-rated texts displayed more varied lexico-grammatical realizations.In these authors' corpus, the expression of authorial stance through attitude seems to be an important factor of success.Nevertheless, compared to other types of stance markersnamely hedges, boosters, self-mention, and engagement markers -, they observe that attitudinal resources are among the least used expressions of interpersonal meanings in both low-and high-rated essays.These findings can be related to those by Crosthwaite and Jiang (2017) in that not only do linguistic realizations of attitude seem to be modulated by EFL writers when writing in English, but they also appear to stem from effective instruction that results in the acquisition of rhetorical conventions of academic EFL writing.Adhering to or neglecting such conventions affects the impression of (in)effectiveness of undergraduate academic texts.
These works have explored and confirmed, to varying degrees and from interrelated perspectives, the possibility to link the use of evaluative attitudinal language to effectiveness in academic EFL writing in different international contexts.The following section presents a description of the methodological approach that was applied in this research in order to explore such relationship in the Mexican context.This will be followed by a report of the results obtained.

Methodology
The purpose of this paper is to analyze realizations of ATTITUDE in undergraduate academic essays in EFL in order to determine if their attitudinal features represent a potential source of text (in)effectiveness beyond lexicogrammatical and structural considerations.This is done through a mixed qualitative and quantitative methodology, comparing the deployment of resources from the three ATTITUDE sub-systems in essays written by Mexican undergraduate students from different instructional levels of the major on English Language and Literature at a public university in Central Mexico.
The major in English Language and Literature consists of eight consecutive semesters.Students entering the first semester must provide evidence of having completed three years of EFL instruction during their high school studies.Additionally, they must pass an exam that determines whether they possess competencies equivalent to band B2 from the Common European Framework.As part of their studies in the major, they study four 16-week language courses during the first four semesters (one course per semester).From the fifth to the eight semesters, their training in the language takes place through specialized content subjects that are taught in English.These subjects include advanced English Literature courses where essay writing is a regular instructional and evaluative practice.

The corpus
The corpus of this work includes 41 essays that were written as final assignments in 16-week English Literature courses by students from different instructional levels: it comprises 15 essays written by students from 1st through 3rd semesterlabeled as 'basic'-, 15 essays from 4th through 6th semesterclassified as 'intermediate' -, and 11 essays from 7th through 8th semesterconsidered as 'advanced' -.The texts were written under the general instruction to analyze one of the literary works studied during one semester, identify its most remarkable literary features, and justify observations about the chosen work.Each essay was evaluated by the teacher of the English Literature course where it was collected and was granted a grade from 5 to 10.During the recollection of the corpus, teachers in charge of the courses reported to have evaluated the essays on the basis of appropriate analysis of literary concepts and features, as well as cohesion, coherence, and grammatical accuracy.No standardized or specially-designed rubric was used, so the results of evaluation resulted from the expert view of teachers according to the academic liberty policy of the university.
For the purposes of this study, the texts were further divided into three subcategories according to the grade they were granted by course teachers: low-graded (essays with notes from 5 to 6), middle-graded (notes from 7 to 8), and high-graded (notes from 9 to 10).Table 6 illustrates the composition and distribution of the corpus.

The analysis
All realizations of ATTITUDE in the corpus were recorded and analyzed taking into account the variables of category, realization, explicitness, and trigger.For the trigger variable, five subject-specific labels were established due to the nature of the analyzed discourse: Author 1 (A1) is marked when the writer evaluates him or herself; Author 2 (A2), when the writer evaluates the author of the literary work he or she discusses in the essay; Character (CH), when the evaluated entity is a character or narrator from the literary work discussed in the essay; Text (TXT), when the writer evaluates elements of the literary work he or she analyzes in the essay, such as atmosphere, style, plot, etc.; and Other (OTH), when the evaluated entity is external to the literary work under analysis, such as literary periods, genres, other works, different authors, etc.Additionally, the authorship of attitudinal evaluation was identified through the variable of appraiser (Martin and White 2005).This variable distinguishes between authorial (the writer is the one who evaluates attitudinally) and non-authorial (the writer attributes an attitudinal evaluation to someone else) realizations of ATTITUDE.
Besides tracking ATTITUDE types in the corpus, the purpose of observing all these realizational variables was to avoid misleading recordings due to potential double codingi.e., invoked realizations of one attitudinal category instantiated by inscribed realizations of a different one.Because a delicate analysis of this type of invoked ATTITUDE is beyond the scope and purpose of this work, when examples of this were identified in the corpus, the inscribed interpretations were considered for the analysis on the basis of prototypical realizations and variables described earlier in Section 2.
Table 7 illustrates attitudinal instantiations from the corpus with the labels of all the variables observed, including an instance of invoked judgement which was recorded as inscribed appreciation (**).In this example, inscribed appreciation (+valuation) is used to evaluate human +veracity indirectly.All instances of ATTITUDE were recorded and quantified in dynamic tables.Proportions of expression of each type of ATTITUDE were calculated by dividing the number of occurrences of each attitudinal paradigm by the total number of clauses in each text and sub-corpus (# of ATTITUDE instantiations/# of clauses).When focus was on ATTITUDE sub-types and realizational variables, proportions were calculated by dividing the number of the corresponding attitudinal paradigm by the total number of ATTITUDE instantiations in each text and sub-corpus.As in previous work on APPRAISAL (Lee 2008;Mei 2007;Valerdi 2016), Chi-squared tests were performed in order to confirm the statistical significance of the findings.This type of test is used to "examine the distribution of data across the categories of our analysis" and "the extent to which the distribution of your observed data varies from the distribution that would be expected if the independent variable had no effect on the dependent variable" (Levon 2010: 78).The students' instructional level (basic, intermediate, and advanced) and perceived effectiveness (low, middle, and high) are independent variables in this study while the expression of ATTITUDE and its variables are dependent variables.
In linguistic studies involving categorical dependent variables, chi-square tests with a resulting p value of <0.05 indicate there is a relationship between the variables (Rasinger 2013).In other words, such tests "tell you that there is at least a 95 % chance that the independent variable does in fact have an effect on the dependent variable" (Levon 2010: 81).The following section presents the results of the study and indicate p values obtained by the chi-square tests.

ATTITUDE in the general corpus and by instructional level
As a general feature, there is a significantly reduced proportion of evaluative language in the whole corpus, with only 18.02 % of the clauses expressing ATTITUDE.This is reflected in each of the instructional level sub-corpora: 19 % in basic essays, 16 % in intermediate essays, and 20 % in advanced essays.Additionally, these subcorpora display appreciation as the most recurrent type of attitudinal positioning, while affect is the least frequent attitudinal paradigm in the corpus.The proportions of the different types of ATTITUDE are statistically significant in all the instructional levels of the texts (Table 8).
When considering the variable of explicitness, the corpus features a progressive reduction of evaluative inscription, which is predominant in basic essays and reduces by 21 % in advanced texts (Table 9).The differences in explicitness among the sub-corpora are highly statistically significant.
Further careful analysis of the explicitness variable reveals that evaluations of affect and appreciation are mainly inscribed in the three groups of texts while resources of judgement tend to be invoked (Table 10).This corresponds directly to the a When a p value is too long to be contained in a single cell, Excel, the program used for quantitative explorations in this work, reports the result with the suffix 'E' followed by an entire number.Such coding means that, for an exact reading of the p value, the decimal point in the result must be moved as many places to the left as indicated by the number following 'E'.In this case, the value of p is statistically significant, as the full result (.) is meaningfully bellow conventional ..Regarding the realizational variable of triggers, authors of intermediate and advanced essays have literary characters, literary features, and external entities related to the works they analyze in their essays as stimuli for evaluative expressions, in that order of saliency and in almost identical proportions.Writers of basic essays, on the other hand, focus primarily on the authors of the works they discuss when positioning themselves attitudinally, but coincide with their intermediate and advanced counterparts in focusing on literary features of their analyzed works secondly, and thirdly on external entities (Table 11).

ATTITUDE by perceived effectiveness
The corpus displays meaningful contrasts in the evaluative features of the essays when their perceived effectiveness is taken into consideration.In Table 12, we can see there is an increasing presence of attitudinal evaluation that progresses from low-graded to middle-graded essays in the basic and intermediate sub-corpora.Such progression, however, does not continue in essays from the advanced sub-corpus, where the distribution of attitudinal realizations contrasts with the other texts and was found to be not statistically significant.Appreciation resources are the most recurrent ATTITUDE type in all low-, middle-, and high-graded essays, followed by judgement and, in much lower proportions, by affect.Additionally, as illustrated in Table 13, proportions of judgement and affect are close in representativeness to each other in all groups of texts, which indicates an interesting regularity related to the degree of effectiveness attributed to the essays by university teachers.
The explicitness variable does not seem to affect the tendencies in the corpus when related to perceived effectiveness (Table 14).Inscribed appreciation and invoked judgement are the most representative choices in low-, middle-, and highgraded essays.Realizations of affect are significantly inscribed with two exceptions to this tendency.First, intermediate middle-graded essays display invoked affect over inscribed realizations by a difference of 20 %.Secondly, advanced low-graded essays are radically different from middle-and high-graded texts from the same instructional level with 100 % of affective realizations being invoked.Despite the clarity of these tendencies, differences in explicitness were not found statistically significant in relation to perceived effectiveness.
Table 15 shows remarkable regularities in the types of triggers that undergraduate writers of two sub-corpora have in mind when expressing attitudinal evaluation of different types.In the intermediate and advanced sub-corpora, all low-, middle-, and high-graded essays display evaluative focus on characters, textual literary features, external elements, the authors of their analyzed works, and the very authors of the essays, in that order of predominance.Interestingly, essays from the basic instructional level differ in every single proportion of trigger choice from the others, with textual literary features as the main attitudinal trigger in all low-, middle, and high-graded essays.The second most representative triggers in lowand middle-graded essays are external elements and characters, respectively, while high-graded essays prioritize authors of their analyzed woks.Finally, low-and attitude in academic EFL writing middle-graded texts coincide in privileging the attitudinal evaluation of the authors of the works they study in the third place.High-graded essays, on the other hand, prioritize their attitudinal stance regarding elements external to their analyzed literary works.These results were found to be highly statistically significant for the basic essays of the corpus, slightly non-significant for the intermediate texts, and markedly non-significant for advanced essays.

Discussion
The results of this research show a meaningful relationship between the use of attitudinal evaluation in undergraduate academic EFL writing and the production of effective texts in the context of the analyzed corpus.In this section, this is shown by discussing the results from general to particular starting with a global scenario of the corpus, then looking at attitudinal features of the texts according to their instructional levels, and finally discussing ATTITUDE in relation to the essays' perceived effectiveness.Following this order, it will be easier to relate the findings to attitudinal features in general undergraduate academic writing and then observe specific features where perceived effectiveness in the texts plays a distinctive role in the corpus of this work.

ATTITUDE in the general corpus
The first relevant observation is the reduced presence of expressions of ATTITUDE in the global corpus, which is a general regularity that previous studies from various contexts have found as a significant feature of academic writing.The low occurrence of attitudinal evaluations in this sort of discourse can be understood in terms of argumentative pertinence.As Hunston states (1999), in academic discourse, only certain things get linguistically evaluated and they do it in specific ways when it is worth it.Additionally, looking back at Petty and Cacioppo's (1986) observations on the effects of subjective language in discourse, the appeal to the affective responses of an audience usually takes place at strategic points where an objective argumentation may lack the necessary persuasive potential.Thus, in discourses where objectivity is highly appreciated, attitudinal evaluations are used in limited proportions and forms, leading to a strategic management of evaluative resources.Such strategic nature in the use of attitudinal evaluations has been previously explored in academic discourse in Spanish and in EFL academic writing.Valerdi (2021) found that ATTITUDE resources occurred in limited proportions due to their usefulness in certain specific types of argument components in postgraduate academic discourses in Spanish.The corpus of this study seems to conform to this strategic management of attitudinal expression.Within the area of EFL academic writing, Jalilifar and Hemmati (2013) concluded that particular dynamics of attitudinal deployment in texts are more determinant than large global amounts of attitudinal language in texts.In line with this, Crosthwaite and Jiang (2017) concluded that effective writing displays limited and careful expression of attitudinal meanings.Considering these antecedents, the general features of evaluative expressions in the corpus of this work seem to respond to academic writing conventions that learners have associated to their use of their target language and, in turn, reflect a general tendency that has been identified as a feature of effective writing in academic contexts.

ATTITUDE types
A second feature the corpus shares with general effective writing is the predominance of realizations of appreciation in more than 50 % of evaluative resources.Such predominance prevails when breaking the corpus into smaller sub-corpora of basic, intermediate and advanced instructional levels.This leaves little room for affective evaluations, which occur in no more than 9 % of attitudinal realizations in all subcorpora.As mentioned before, only certain things are evaluated in academic texts, and they are in certain ways only.This is confirmed when comparing these findings with those of previous works on APPRAISAL in different contexts.Zhang and Cheung (2018), for instance, found appreciation as the most realized type of ATTITUDE in articles on Second Language Writing, which they see as the result of a strategic evaluation which focuses on the value of things in order to objectify observations that are subjective in nature.Lee (2015) observes similar tendencies in undergraduate essays and emphasizes the importance of evaluating things related to the subject matter in the production of academic texts.According to these interpretations, prioritizing the appraisal of discipline-related things is central in academic arguments, where evaluation tends to be more objective than it would were it based on observations about human behavior or emotional responses of writers.

Attitudinal explicitness
The inscription and invocation of attitudinal meanings display one more regularity of the corpus that remains constant in the basic, intermediate, and advanced subcorpora.From a general perspective, attitudinal evaluations are predominantly inscribed, making their attitudinal stance accessible in terms of the effort their identification and interpretation demand from the reader (Halliday 2002).
Nonetheless, such inscribing tendency changes when looking at each attitudinal category; realizations of judgement tend to be invoked in more than 65 % of occurrences.
These features carry interesting implications in terms of the degrees of explicitness with which EFL academic undergraduate authors deploy their evaluations.First, the global tendency to inscribe evaluations coincides with Hood and Martin's (2005) observation that academic writers do not usually invoke ATTITUDE when constructing arguments around their work.The features of the corpus of this work confirm the validity of that observation in EFL academic writing.Regarding the contrasting realizations of judgement, previous research has shown how evaluations focused on people and their behavior tend to be managed more carefully in academic contexts than those triggered by things, taking the form of invoked attitudinal resources (Hood and Martin 2005;Valerdi 2016).Apparently, the authors of the corpus behave more freely or confidently when evaluating things related to the subject matter of their work and their own personal impressions through appreciation and affect than when dealing with humanor humanizedtriggers, in which case they tend to proceed more cautiously.It seems plausible to say these features correspond to generally effective evaluative dynamics in academic writing.Even though proportional differences between inscribed and invoked affect are not statistically significant (Table 10), the regularity of their plain contrast and the significance of differences in judgement and appreciation stand as a remarkable feature of the corpus.

Triggers of attitudinal evaluation
The tendencies of the most recurrent triggers of attitudinal evaluation in the corpus are another area of significant regularity across instructional levels.The vast majority of ATTITUDE realizations is triggered by elements inherently associated to the literary works the authors analyzed in their essays; in the overall corpus, the most recurrent triggers are literary characters, followed by textual features from the realm of literary studies, and entities and concepts external to literary works in second and third place, respectively.These results seem to be a direct consequence of the instructions authors were given to analyze a literary work by identifying its most remarkable features as instances of literary genres and to justify their observations.Here, the notion of justification is key; since the authors were expected to justify their observations about literary features, it would seem natural to expect arguments founded on argumentative conclusions and supporting ideas focused on such things as characters, textual features and the contexts around literary works.Such was the case of the corpus of academic essays analyzed by Myskow and Ono (2018), who found that attitudinal resources were used as part of the two central components of arguments, namely conclusions and supporting ideas, connected to subject matterrelated triggers.Taking into consideration Myskow and Ono's experience, together with Valerdi (2016) findings on the realizations of attitudinal evaluations directly connected to specific argument components, the statistically significant features of the triggers variable in this corpus confirm the centrality of attitudinal evaluation on the elaboration of academic arguments.

ATTITUDE by instructional level
The corpus displays significant variations if the results are seen from the perspective of the three levels of instruction of the authors of the texts.There is a clear tendency for uses of affect and appreciation to decrease progressively from basic through advanced essays, while realizations of judgement increase from 13 % to 40 % (Table 8).This could be interpreted as the direct consequence of the authors' choice to evaluate characters from the works they studied and, as a result, it further supports the strategic nature of evaluative choices in the corpus; if any type of ATTITUDE was to remain present at different instructional levels, it was the one triggered by humanlike entities.Additionally, despite the general reduction of inscribed evaluations in all three sub-corpora (Table 9), expressions of judgement developed on the opposite direction, reducing their invoked realizations significantly (Table 10).These observations suggest interesting evaluative dynamics on the corpus; while there is a general caution on the part of authors to evaluate human and human-like entities in their texts, there also seems to be certain development of a notion of discursive authority (Poynton 1985, as cited in Valerdi 2016) allowing writers to express their stance regarding human triggers in more open terms through inscribed judgmental evaluations as they gain more linguistic and disciplinary experience.
The relationship between evaluative language and disciplinary experience can also be seen in the triggers on which authors focus their ATTITUDE realizations.Writers of basic essays contrast with writers of intermediate and advanced texts in their evaluative focus on authors of literary works as the third most relevant triggers (Table 11).Clearly, the interpretation of writers of basic essays regarding what elements are central in literary analysis is particular.Considering previous observations about the relevance of prioritizing arguments around the value of things to objectify evaluations (Zhang and Cheung 2018), it seems plausible to conclude that, in intermediate and advanced texts, the less central consideration of human elements different from literary characters is the result of the development of a more specialized selection of elements worthy of literary discussion.These findings further support what has been observed by Candarli et al. (2015), Lee and Deakin (2016) and Crosthwaite and Jiang (2017); attitudinal choices seem to be modulated by EFL writers when writing in English and, in parallel with this, they stem from disciplinary instruction which leads to the acquisition of academic EFL writing rhetorical conventions.On the grounds of these observations, it is clear how previously discussed general features of effective writing develop in the corpus as essays progress towards more experienced writing.

ATTITUDE in relation to perceived effectiveness
The results of the analysis evidence a tendency in high-graded essays to display significantly greater proportions of ATTITUDE than texts graded as low and middle (Table 12).The fact that the advanced sub-corpus displays an irregular distributionwith middle-graded essays presenting the most ATTITUDEcould be explained by the acquisition of more regularly applied conventions by advanced writers resulting in closer proportions between these groups of texts.Further explorations of advanced essays in this and similar contexts are needed to confirm this.Still, even though the p value of 0.11234221 indicates results might be different for this sub-corpus if this research were replicated, in the actuality of these results low-graded essays display the least instances of attitudinal evaluation.
It is remarkable to notice the significant predominance of appreciation over judgement and affect stands when effectiveness is considered.This indicates the development of ATTITUDE-related interpersonal conventions along basic, intermediate, and advanced instructional levels has been shaped by pedagogical dynamics that favor the observed tendencies via both instruction and academic evaluation: what teachers disseminate and evaluate as effective motivates what learners use to shape their discourse in this type of writing as they move towards more specialized levels of literary analysis.It is worth remembering that, according to what they reported, teachers who assessed the essays did not consider attitudinal expressions in their evaluation criteria.Then, in line with Lee's analysis (2008), we can conclude that teachers might have perceived the management of evaluative language, which directly relates to writers' stance and voice, as part of the appropriate analysis of literary components and features they sought; ATTITUDE is central to effectiveness in the corpus of this work.
Similar findings in previous research indicate the latter interpretation may also apply to other contexts; Jalilifar and Hemmati (2013) found successful EFL texts by Iranian writers displayed significantly more frequent instances of appreciation than judgement and affect, just as Myskow and Ono (2018) did in their essays by Japanese students.Moreover, similarly to this research, patterns of attitudinal inscription and invocation in the corpora analyzed by those authors can be significantly related to degrees of (in)effectiveness.In the corpus of this work, inscription of affect and appreciation contrasts with invocation of judgement in low-, middle-, and high-graded essayssave for minor variations in affect, the least representative category, in the basic and intermediate sub-corpora.Although these contrasts may require further exploration to confirm statistical significance, their plain tendencies and their consistency with the significant dynamics by instructional level confirm the role of attitudinal meanings in the shaping of undergraduate writers' positioning and, most importantly, a positive relationship between realizations of ATTITUDE and the effectiveness of undergraduate EFL academic writing instantiated by the corpus.
Such relationship is further evidenced by the development of attitudinal triggers selection patterns.As it was previously described, in both intermediate and advanced essays, characters, textual features and external literary elements are the most representative triggers in low-, middle-, and high-graded essays.In the basic sub-corpus, however, the very authors of the essays figure as either the second (lowand middle-graded texts) or first (high-graded-texts) choices.The contrast this represents can be explained, once more, on the basis of instruction and academic evaluation; although attitudinal stance focused on the very authors of the essays conflicts with the general features of the corpus, the fact that the other two main trigger types in basic essays remain within what can be seen as disciplinarily central in the general scenario may have allowed for certain tolerance on the part of teachers.Although having access to feedback received by the writers of the basic essays would be the only way to confirm this, it should be enough to notice that trigger selection in relation to text effectiveness takes a clear orientation in the rest of the corpus, thus corroborating the role of attitudinal evaluations in undergraduate academic writing instantiated by the corpus.
In the light of these results, a positive relationship between the expression of attitudinal meanings and academic texts (in)effectiveness has been confirmed.What is more, the writing conventions followed by the authors of the corpus of this work suggest that attitudinal features privileged by university teachers in basic levels of instruction determine the conventions adopted in effective texts at more advanced levels, even when attitudinal dynamics do not seem to have been explicitly prioritized.Apparently, attitudinal dynamics in the context of the corpus have developed from complex influences which go beyond mere linguistic and disciplinary instruction.These observations relate to Chitez and Kruse's (2012) considerations about writing cultures: further extra disciplinary features related to writing cultures shape what could be considered as effective or successful in writing academically.These include learners' class experience before and during university education, contact with diverse curricular arrangements, university-specific organizational structures, national writing cultures, and differences among languages.As evidenced by the findings of this work, these factors represent important aspects to consider in the design of academic writing programs and discipline-specific writing courses as these may have to take into consideration the features of their participants' target audiences, including potential expectations determined by their own linguistic, cultural, and even institutional backgrounds.
All these considerations have a particular centrality in programs focused on training future professionals in English Language and Literature for, as it has been shown, even when interpersonal dynamics involving ATTITUDE conform to similar general tendencies in the whole corpus of this study, they vary significantly at more delicate degrees of analysis such as instructional levels and realizational variables.Additionally, exploring such delicacy in variation, which can only be done by analyzing the academic discourse of speakers from different instructional levels within the same undergraduate community, may also serve as a valuable source of information for pedagogical actions aiming to guide undergraduate writers in the production of effective texts in other disciplines.

Conclusions
As the findings of this work have shown, the construction of interpersonal relationships through the use of linguistic resources of ATTITUDE plays an important role in the perceived effectiveness of undergraduate EFL academic writing at different levels of experience and disciplinary instruction, which indicates the importance of broadening assessment criteria of academic writing beyond the constraints of lexicogrammatical accuracy and text structure, as well as pertinent observations about attitudinal positioning in instructional processes, in favor of deeper interpersonal metalinguistic awareness in both students and writing instructors.In order to face these instructional challenges, it is important to emphasize the results also show that the influence of attitudinal language in the (in)effectiveness of academic texts is furtherly determined by a complex series of contextual factors including discursive authority, writing experience, discipline-specific focus objects, and the multi-faceted writing-related cultural background of writers.
These conclusions are supported by six main findings that have been reported in this paper: 1) In adherence to general academic writing, the corpus displays a significantly low occurrence of ATTITUDE resources, which results from a strategic deployment of interpersonal resources by writers as to what is evaluated in their texts and in what ways; 2) General attitudinal features in the corpus privilege appreciation as the most frequent type of ATTITUDE and position affect as the least deployed evaluative resource, which reflects the writers' prioritizing of evaluations of discipline-related things to project an objective-like positioning; 3) Realizations of judgement throughout the corpus are mainly invoked, signaling the adoption of particular evaluation strategies around human or human-like triggers; 4) Basic essays feature greater use of attitudinal resources than intermediate and advanced essays, suggesting a progressive development of objective-like writing dynamics; 5) Overall, high-graded essays display more frequent expressions of ATTITUDE than middle-and low-graded essays, suggesting that the expression of ATTITUDE is not simply reduced in learners' discourse as they move forward in their academic training, but according to conventions that favor certain types of strategic attitudinal positioning in texts; 6) In general, the realizational variables of ATTITUDE category, explicitness and trigger selection show significantly matching tendencies in highgraded essays in the basic, intermediate and advanced sub-corpora, confirming the central role of the strategic use of affect, judgement and appreciation resources in terms of their realizational variables in the production of effective texts, even when writing conventions develop towards reduced proportions of ATTITUDE.Those findings which were not found to be statistically significant might need further exploration, perhaps on the basis of a larger corpus, in order to refine their implications from a statistical perspective.Still, it is worth remembering statistical interpretations, either descriptive or inferential, always leave room for open possibilities.
Regarding the latter point, interpretations of the reported results need to take into consideration that a relevant limitation of this study is the size of its corpus.Although the results of the analysis demonstrate consistent and mostly statistically significant features, their full implications apply to the context of the academic community where the essays were collected.In order to obtain more generalizable findings with more predictive potential in relation to the Mexican context, future work might need to replicate the study on the basis of a larger and more varied corpus.
While the scope of the results of this work is limited to the boundaries of the context of this research, the implications of its findings represent meaningful contributions that complement previous research on the workings of APPRAISAL resources of ATTITUDE in EFL academic writing.Most importantly, this work contributes to further understanding of the workings of evaluative language in academic instruction in the context of Mexico, where pedagogical developments from the systemic functional perspective of APPRAISAL theory are still in their early stages.It is ideal that the focus of future work seek to implement research findings in the development of concrete instructional tools which improve the scenario of effective interpersonal positioning for undergraduate authors seeking to take part in international disciplinary dialogue through academic English.

Table  :
Instances of the four categories of affect.

Table  :
Instances of the five categories of judgement.

Table  :
Instances of the three categories of appreciation.

Table  :
Invoked realizations of ATTITUDE.

Table  :
Instantiations of ATTITUDE with realization variables.

Table  :
Composition and distribution of the corpus and sub-corpora.

Table  :
Instantiations of ATTITUDE with labels of five realizational variables.

Table  :
Proportions of realization per ATTITUDE sub-system by instructional level.

Table  :
Proportions of inscribed and invoked ATTITUDE by instructional level.
and mental processes for affect; attributes and epithets for appreciation; and processes, nouns and nominalizations for judgement.The proportions of explicitness were found statistically significant for judgement and appreciation only.

Table  :
Proportions of invocation and inscription of ATTITUDE types by instructional level.Predominant tendencies are highlighted.

Table  :
Triggers of attitudinal evaluations by instructional level.The three most representative types of triggers in each sub-corpus are highlighted.

Table  :
Proportions of ATTITUDE in relation to grades granted by university teachers.

Table  :
Proportions of ATTITUDE types in relation to grades granted by university teachers.

Table  :
Proportions of invocation and inscription of ATTITUDE types in relation to grades granted by university teachers.Predominant tendencies are highlighted.

Table  :
Triggers of attitudinal evaluation in all sub-corpora in relation to grades granted by university teachers.The three most representative types of triggers in each sub-corpus are highlighted.