English for Specific Purposes: Traditions, Trends, Directions

English for Specific Purposes (ESP) began around fifty years ago as a result of pressing worldwide demands for fast-paced language training in occupational and professional settings, rapid revolutions in theoretical linguistics, and burgeoning pressures on schools and educators to focus on, and to be responsive to, learners’ needs. It started within the field of English Language Teaching (ELT) but has gradually established itself as an autonomous subfield of Applied Linguistics (AL). This paper will review the origins, evolution, and status quo of ESP, and then predict the future directions of this important field. The theoretical, analytical, and methodological evolutions of ESP are reviewed, the positions of genre analysis, target language use situation analysis, and context in ESP are described, the ‘just-in-case’ EAP and ‘just-in-time’ EOP approaches are compared, and the text-first and context-first approaches to discourse structure analysis are compared. The paper predicts that ESP will adopt a wide-angled epistemological stance to survey the (a) discursive, (b) generic, (c) social, and (d) organizational structures of specialized texts and discourses, as well as those of texts and discourses simplified for the popularization of science, in a systematic and contextualized manner. ESP practitioners are also warned about the potential threats of teaching genres of power within ESP.


INTRODUCTION
About fifty years ago, English for specific purposes (ESP) started within the discipline of English Language Teaching (ELT) with the aim of helping international students with their academic writing tasks at universities where English is the medium of education. A less important aim was to help scholars from non-English countries with their publications in English-medium journals (Johns, 2013;Johns & Dudley-Evans, 1991). It gradually grew out of the pressing demands of a fast-developing postwar world in which satellite communication, rapidly-growing international trade, multinational companies, digital technology, etc. made access to, and competence in, an academic/business/trade lingua franca inevitable. In that context, the developments of language teaching methodology in the 1960s and the 1970s (specifically the emergence of Communicative Language Teaching (CLT) and the Notional Functional approach) made ESP a plausible and cost-effective way of teaching English to non-TESOL learners.
In her discussion of the origins of ESP, Robinson (1991) argued that it has emerged as a result of (a) worldwide demands, (b) a revolution in linguistics, and (c) focus on the learner. Johns and Salmani Nodoushan (2015) suggested that the learner as a psycho-socio being was traditionally much less the focus in ESP than other approaches to language teaching. They, however, agree that focus on learners makes sense if it is tantamount to preparing them for their later responsibilities and the demands that they may have to respond to in their future professions. Today, ESP is a well-established subfield of applied linguistics (some would say it is an independent field by and in itself), and its evolution is still going on. In this paper, I will (1) provide a brief review of the past few decades of ESP, (2) distinguish between 'just in case' and 'just in time' approaches to the teaching of ESP, (3) discuss Content and Language Integrated Learning as well as Content-Based Learning, and (4) outline the future directions of ESP.

ESP vis-à-vis EGP
Apparently, the most logical differentiation that one might make would be between English for General Purposes (EGP) and English for Specific Purposes (ESP); I personally have always felt uneasy with this dichotomy, though; by the same token that Mackay and Mountford (1978) rejected ERL as a specific form of language, I have always considered ESP as the practical application of language to a specific 'use' context-not as a self-sustained autonomous language in and of itself (Salmani Nodoushan, 2002). Perhaps an easier way to think of these two terms is to think about the difference between 'driving' and 'cars'. Driving is a kind of competence, but cars are what can be operated by a person who has that competence. Anyone who knows driving will be able to drive any brand of cars-of course, with some adjustments. By analogy, anyone who knows English will be able to learn how to 'use' it in specific contexts. What they need is training in the relevant genres and/or discourses, not learning a language other than English; they need to adjust their 'language use' to the requirements of the specific genre that they need in the target language use (TLU) situation (i.e., familiarity with rhetorical moves, metadiscourse, register, and so forth). In other words, EGP is discipline-independent language use whereas ESP is disciplinespecific language use. Hutchinson and Waters (1987), too, argued that EGP and ESP do not differ in theory, but they differ a lot in practice. They argue that "ESP is an approach to language teaching in which all decisions as to content and method are based on the learner's reason for learning" (p. 19)-cf., Dudley-Evans (1997).
Part of the claim about EGP and ESP being two distinct fields comes from misinterpretations of Swales (1990) in the treatment of genre. Swales' focus on, and evolving thinking about, 'purpose' led him to conceive what we know as genre analysis today (Salmani Nodoushan, 2011Salmani Nodoushan & Khakbaz, 2011Salmani Nodoushan & Montazeran, 2012). Both 'purpose' and 'genre' are major concepts in genre analysis (Bhatia & Salmani Nodoushan, 2015;Johns, 2015;Swales, 1990), but Swales' treatment of genre analysis never claimed EGP and ESP to be two different languages, but two intertwined competencies. The term 'genre', à la Swales (1990), pertains to (a) purpose and (b) text prototypicality.
Although in his 1990 volume, Swales was quite specific about the relationships between purpose and genre, he is no longer quite so sure (Johns & Salmani Nodoushan, 2015). Back in the 1990s, Swales identified 'communicative purpose' as the key feature that could tell if a given text was an instance of a given genre. In the 21st century, however, he has changed his mind and has argued that (a) texts are not necessarily genre-specific, (b) any given genre may serve multiple purposes at the same time, and (c) these purposes are learner-or user-specific (Askehave & Swales, 2001). Earlier, Swales and Rogers (1995) had argued that linguistic and/or rhetorical similarities among instances of a given genre do not guarantee that they will have to serve the same communicative purpose. To complicate this picture even further, Swales (2004) argued that communicative purpose is not static; it may evolve through time (i.e., change, shrink or expand). Moreover, texts belonging in the same genre category may serve different communicative purposes across cultures-i.e., text and communicative purpose are culture-bound.
This change of perspective has resulted in the emergence of a rich repertoire of professional terms in genre analysis. Swales himself coined several terms (e.g., genre constellations, genre hierarchies, genre chains, genre sets, genre networks, subgenre, etc.) of which the discussion is beyond the scope of this paper (For a thorough discussion, please see Salmani Nodoushan (2011). By the same token, Bhatia (2004) coined 'genre colonies' to refer to genres that serve the same communicative purpose across different professions, disciplines, and contexts of use. Johns (2015), too, argued that a single text from a genre is likely to serve several purposes, for both its writer and its audiences; she further argued that it can no longer be claimed that there is a one to one correspondence between a given text in a given genre and a given purpose. As such, ESP and EGP cannot be differentiated in linguistic terms but in communicative and/or functional terms-or in terms of genre, text types, task types, purpose, contexts, TLU situations, and so forth.
Nevertheless, this is not specific to ESP, and one can safely argue that all instances of language use are 'specific' in that each language use situation requires certain linguistic codes. As Hymes (1974) has rightly argued, any competent language user evaluates the components of any given speech situation (i.e., SPEAKING) and chooses an appropriate code to serve that situation. In the acronym SPEAKING, (1) S refers to the setting (i.e., the time, place, physical circumstances, and psychological setting or scene); (2) P refers to participants (i.e., speaker, addressor, hearer, and addressee); (3) E refers to the ends (i.e., purpose, outcomes, and goals); (4) A refers to act sequences (i.e., message content and message form); (5) K refers to keys (i.e., manner/spirit in which something is said); (6) I refers to instrumentalities (i.e., channels and forms); (7) N refers to norms (i.e., norms of interaction and interpretation); and (8) G refers to genres (i.e., categories of communication)-cf., Salmani Nodoushan (1995).
There are, of course, people-like Chapelle (1998) and Douglas (2000)-who would argue that Communicative Language Ability (CLA) (cf., Bachman, 1990;Canale & Swain, 1980) is different from Language for Specific Academic Purposes (LSAP) ability. The concept of "interactionist view" of construct definition was first broached by Chapelle (1998) who believed taking (a) language users' traits and (b) context features into account but ignoring the interaction between the two-and overlooking the probable synthesis that they can achieve together-is artless simplicity. Their interaction, à la Chapelle, can change the quality of each one of them so that trait components can longer remain absolute and context-independent, and contextual features, in turn, will have to be defined with reference to their impact on trait components (Chapelle, 1998). She, therefore, called for a theory of "how the context of a particular situation within a broader context of culture, constrains the linguistic choices a language user can make during a linguistic performance" (p. 15); she did not provide any evidence to support the psychological reality of her claims, though. This is what Douglas (2000) expatiated upon to hypothesize the existence of what he called Language for Specific Purposes (LSP) knowledge, the nature of which may differ from other forms of language knowledge. According to Douglas (2000), LSP ability can get engaged in communicative tasks only when it interacts with, and is aided by, external context. This claim suggests that ESP competence cannot be the same as Bachman (1990) communicative language ability. Like Chapelle (1998), Douglas (2000) did not provide any evidence in support of the psychological reality of his claims either. Johns and Salmani Nodoushan (2015) agreed that we are partly constructed by our experiences and environments and that, if anything, LSP ability might be a social reality (Douglas, 2000). They also agreed with Swales (1990) on 'enduring conceptions' in ESP (i.e., authenticity, research-base, learning/methodology, need, and language/text) but added context to this list. In their consideration of the role of context in ESP, Johns and Salmani Nodoushan (2015) agreed with claims of the 1990s' that ESP is protean in that it is responsive to developments in (a) language, (b) pedagogy, and (c) content studies. Nevertheless, they did emphasize that ESP needs to be even more protean today to be sensitive to (a) the context of teaching and learning, (b) time constraints, and (c) the variety of content as well as professional studies in which learners are involved.

ESP IN EARLY 21st CENTURY
By default, any field of science-including any subfield of applied linguistics-develops in at least three domains: (a) theoretical, (b) methodological, and (c) analytical (Salmani Nodoushan, 2016; 2020a).

Theoretical Developments
The evolution of ESP in the 20th century, especially after Chapelle (1998) proposed her "interactionist view" on the definition of the construct of ESP ability, was expected to promise attempts on the part of ESP scholars at theory making and development, but this line of scholarly activity has unfortunately remained the Cinderella of ESP-except for LSPA (Douglas, 2000). Although Hyland (2005), and more recently Bhatia (2017), also proposed their views on-or theories of-'metadiscourse' and 'critical genre analysis' respectively, their frameworks are analytical, but not that theoretical. In other words, the main developments of ESP in the 21st century have remained in the domains of analysis and methodology. Nevertheless, the notion of 'discourse community' by Swales (1990) has profoundly influenced ESP research in the last two decades (See also Donadio, 2019).
It is no surprise that the 'theoretical' domain in ESP has not received enough attention. ESP has mainly been approached from a pedagogical perspective partly because the main concerns of its practitioners have usually been the teaching methods and materials of which the relevance can only be judged in the light of their direct bearing on (a) the didactic process and (b) its outputs (Fiorito, 2019). Nevertheless, I still have my own controversial belief expressed in the 'cars-and-driving' metaphor in the previous section-i.e., ESP is just the practical application of language to a specific 'use' context, but not a self-sustained autonomous type of language in and of itselfunless we may want to adopt a sociolinguistic perspective on ESP and argue in favor a pidginized, creolized, or vernacular version of English (cf., Andersen, 1983;Bickerton, 1977;Schumann, 1978aSchumann, , 1978b spoken by a professional minority, but note that in that case we can only speak of the social/sociolinguistic reality of ESP, but not its psychological reality (cf., Johns & Salmani Nodoushan, 2015). According to Fiorito (2019), the need for a research foundation for the teaching and learning of ESP has come to academics' and ESP teachers' focal attention in recent years. In lieu of walking in the footsteps of Chapelle (1998) and Douglas (2000), the majority of ESP teachers and practitioners in the past two decades have opted out of theoretical issues and turned to methodological and analytical studies which comprise a variety of research focused on metadiscourse, genres, corpora, and so forth.

Analytical Developments
In this connection, it should be noted that earlier reviews of the history of ESP had identified five major transformations: (a) register analysis, (b) rhetorical discourse analysis, (c) target language use situation analysis, (d) skills and strategies, and (e) a learning-centered approach (Salmani Nodoushan, 2002). Johns (2013) emphasizes the fact that genre analysis by Swales (1990) has had a huge impact on how rhetorical discourse analysis has been approached in ESP research in the 21st century (See also Johns & Salmani Nodoushan, 2015). As such, the notion of genre has become a quite significant force in ESP research and teaching, and many scholars have brought it to bear on their research activities (Hyland, 2004a(Hyland, , 2004bHyland, 2007;Paltridge, 1997Paltridge, , 2001Paltridge, , 2007Paltridge, , 2013Swales, 1990Swales, , 2004. Many of the genre-based studies that have been conducted in ESP in the past two decades or so rely heavily on Create A Research Space (CARS) model (Swales, 1981) to describe, and account for, the typical discourse and/or rhetorical structures of academic texts-mainly with a pedagogic purpose (cf., Celce-Murcia & Larsen-Freeman, 1999). The genre-based approach to the analysis of discourse structures has not remained within the confines of ESP; it has found its way to professional genres (e.g., the popularization of science for media, TED talks, etc.), and many professional-genre studies are informed by Bhatia (1993Bhatia ( , 2004. As Bhatia (2004), Flowerdew (2002), Swales (2004), and Askehave and Swales (2001) have all noticed, a review of the studies conducted on the discourse structure of any given genre reveals that researchers have followed either of the two approaches: (1) the text-first approach, or (2) the context-first approach. In the former, text/corpus selection is done first and then the researcher exploits the text to find its (meta)discursive, generic and rhetorical patterns. In the latter, texts are analyzed in the light of the features of their already-established and well-described contexts. The former deepens our knowledge of the discourse and linguistic features of texts in an academic discipline or a profession; the latter tells us to what extent a given text complies with or deviates from the context to which it relates, and deepens our understanding of the features of context. The context-first approach has motivated genre studies in ESP to move beyond linguistic descriptions, and researchers working in this camp are more interested in finding (1) why a given genre is shaped as it is, and (2) how it achieves its goals. As Paltridge (2013) noted, there are still researchers (e.g., Martin & Rose, 2008) who prefer the text-first approach, but they draw on Halliday's systemic functional framework (Halliday, 1985(Halliday, , 1994Halliday & Matthiessen, 2004) for text analysis, and use 'schematic structure' rather than 'genre' to refer to the discourse structure of texts (See also Gonzalez, 2016;Heberle & Morgado, 2016;Hood & Lander, 2016;Jesus et al., 2016;Macken-Horarik & Sandiford, 2016;McQueen, 2016;Oteiza & Pinuer, 2016;Rivas & Germani, 2016;Shum et al., 2016;Tehseem, 2016;Zuppa & Rezzano, 2016).
Another line of analytical development owes much to innovations and advancements in computer software technology and the Internet which have made (a) the storage and analysis of large sets of texts and corpora both possible and easy, and (b) the generalizability of findings more precise. A few examples of early corpus-based studies are the ones conducted by Biber and Conrad (2009), and Flowerdew (2011. This new computer-assisted analytical approach, sometimes called corpus linguistics, has revolutionized ESP. It has brought the notion of 'specific' language to the fore again, and we can now hope to be really 'authentic' and 'precise' at the level of language (Johns & Salmani Nodoushan, 2015). Corpus linguistics has made it inevitable for us to engage multi-media technologies in our research, and such technologies will eventually (a) affect our research and pedagogies in the target situation at the very basic level and (b) determine how we get our students ready for those eventualities (Johns & Salmani Nodoushan, 2015).
Advanced technology has also made it possible for ESP researchers to design and conduct cross-cultural and cross-disciplinary studies. Connor, for instance, has summarized cross-cultural examinations of the CARS model by Swales (1990) under the heading 'Contrastive Rhetoric' (Connor, 1996) and the heading 'intercultural rhetoric' (Connor, 2011). Studies of metadiscourse, too, have benefited a lot from advanced technology. Concordance software packages have made it possible for researchers to analyze metadiscursive features (e.g., hedging) in ideational, textual or interpersonal metadiscourse. As such, digital technologies partly determine students' learning outcomes (i.e., what students need to know, understand, and do in TLUs) and their workload (i.e., how much time they have or need to achieve learning outcomes (Anderson, 2017)).

Methodological Developments
As for methodological developments, ESP has witnessed three major approaches to pedagogy in the past two decades: (1) content-based instruction (CBI), (2) content and language integrated learning (CLIL), and (3) genre-based instruction (GBI)some would also include task-based instruction (TBI) here, though.

CBI
CBI dates back to 1965 and is often associated with language immersion programs in post-colonial Canada. In their definition of CBI, Brinton et al. (1989) argued that CBI integrates 'language teaching' with 'particular content' to pave the way for "the concurrent teaching of academic subject matter and second language skills" (p. 2). This implies that target language is not an object of study but a medium for teaching and learning subject matter content. However, they seem to have confused CBI and CLIL. Genesee (1994) perspective on CBI is more precise.
In response to the question of what can qualify as 'content' in CBI, Genesee (1994) suggested that, when teaching general language skills is at stake, content does not need to be academic in nature but can include "any topic, theme, or non-language issue of interest or importance to the learners" (p. 3) provided that it can impart appropriate knowledge and command of those skills to them. Along the same lines, Met (1991) has argued that content in CBI will have to be cognitively engaging and demanding for students to be able to foster the required skills and abilities in them. Likewise, Eskey (1997) has claimed that CBI does not teach "the content itself but some form of the discourse of that content" (p. 139). As such, a CBI teacher is expected to "acculturate students to the relevant discourse communities" and a CBI student is expected "to become acculturated to those communities" (Eskey, 1997, p. 140).
The important question that needs a convincing answer is: What is the rationale for content-based instruction? To answer this question, it should be emphasized that CBI has indeed received support from at least four different domains: (1) second language acquisition (SLA) research, (2) research on instructional strategies, (3) educational and cognitive psychology, and (4) research on program outcomes (cf., (Grabe & Stoller, 1997). First, SLA research has shown that natural language acquisition/learning cannot be divorced from meaning, and that meaning requires a context for meaningful communication where language is learnt in use (Curtain, 1995;Lightbown & Spada, 1993;Met, 1991;Wells, 1994). Lightbown and Spada (1993) have also showed that that negotiation of meaning promoted by CBI can enhance language acquisition. Likewise, Krashen (1982Krashen ( , 1985 has shown that CBI engages comprehensible input to facilitate SLA. Lyster (1987) and Swain (1985) have also noticed that form-focused content in CBI enhances SLA. By the same token, Cummins (1981) cognitive academic language proficiency (CALP) has supported CBI (cf., Byrnes, 2000). Other studies on SLA that have supported CBI include Genesee (1994), andLantolf andAppel (1994). Second, research on instructional strategies has also shown that CBI (a) is suitable for cooperative learning (Crandall, 1993;Slavin, 1995), (b) promotes learning strategies and develops strategic competence (O'Malley & Chamot, 1990), (c) fosters learning strategies and rich thinking and language skills in learners (Curtain, 1995;Met, 1991), and (d) results in improved language abilities, greater content knowledge, and higher motivation through extensive reading content (Elley, 1991). Third, CBI has also been supported by research from cognitive and educational psychology. The cognitive learning theory by Anderson (1990Anderson ( , 1993 with its three transitional stages (i.e., cognitive, associative, and autonomous) has shown that CBI has merits in that it (a) facilitates the process of 'proceduralization' of knowledge, and (b) leads to deeper processing, greater subsumable learning, and better recall. Likewise, Singer (1990) has shown that thematically organized information (which is a merit of CBI) results in improved learning. Byrnes (2000), too, has argued that CBI is more cognitively engaging and fosters a wider range of discourse skills in students than other instructional methods. Csikszentmihalyi (1997) has also suggested that content-based activities turn up into 'flow experiences' that lead to optimal mastery. Finally, research on program outcomes has shown that CBI (1) leads to better language and content learning, higher levels of motivation and interest, and greater employment opportunities; (2) is more tailored to immersion and bilingual programs, post-secondary ESL/FL contexts, and FLAC programs; (3) paves the way for greater flexibility in curriculum and class activities; (4) allows greater adjustment to learners' needs and interests; and (5) bridges the gap between basic language study and advanced cultural and literature studies (Grabe & Stoller, 1997).

CLIL
It should be noted here that CBI and content and language integrated learning (CLIL) differ in subtle ways although people who are less versed in the field of ESP may take them as alternative terms referring to the same methodology. It was stated above that CBI has its roots in language immersion programs that were the vogue in post-colonial Canada, and that its focus is teaching language-but not content; According to Genesee (1994), "any topic, theme, or non-language issue of interest or importance to the learners" (p. 3) can be used in CBI classrooms since the aim is to teach language. By way of contrast, CLIL is focused on the teaching of both language and particular academic subjects. In a CLIL classroom, academic subjects such as biology, philosophy, history, etc. are taught through a foreign language. As such, CLIL can be defined as a teaching methodology that dovetails language and academic content and teaches them in tandem so that students can learn both. According to Marsh (2002), "CLIL refers to situations where subjects, or parts of subjects, are taught through a foreign language with dual-focused aims, namely the learning of content, and the simultaneous learning of a foreign language" (p. 2, italics mine). Likewise, Coyle et al. (2010, p. 1) defined CLIL as a "dual focused educational approach in which an additional language is used for the learning and teaching of both content and language". Some would say that CLIL is an invention from the noughties (2000-2009), but it has been around at least from the 1960s. Nevertheless, it has grown exponentially throughout the world, and specifically in Europe, in the past two decades, partly due to its taken-for-granted potential to foster gains in both language teaching and subjectmatter education. Perhaps it is more realistic to assume that universities, schools and companies have welcomed CLIL due to its cost-cutting and/or bureaucratic potentials rather than its educational value. From an educational perspective, CLIL creates greater challenges for both students and teachers. Rarely is it possible for universities, schools and companies to hire or tenure a teacher who is competent in both the language and the content that are to be taught. Likewise, simultaneous focus on language and content is quite demanding for students as it creates extra cognitive workload for them. Nevertheless, the ardent proponents of CLIL have argued that, in spite of its being a great challenge, it is "more motivating and authentic for students and teachers" (Marsh, 2002, p. 72). Such claims have not been empirically researched enough yet; only a few studies have been conducted on the (dis)advantages of CLIL and the specific pedagogical practices of its teachers (Dalton-Puffer & Smit, 2013;Pérez-Cañado, 2012)-added that their findings are, of course, controversial.
CLIL should be distinguished from not only CBI but also any other form of immersion program or bilingual education. For one thing, CLIL is a kind of principled pragmatism (cf., Salmani Nodoushan, 2006) and an eclectic "type of instruction that fuses the best of subject matter and language teaching pedagogies" (Morton, 2010, p. 97). Secondly, the lingua franca or the foreign language that is used in a CLIL classroom is not the language spoken outside of it. Thirdly, CLIL students are expected to have already acquired and mastered literacy skills in their mother tongues before they are exposed to CLIL. Moreover, CLIL lessons are essentially regular and timetabled within the school curriculum. Finally, the teachers teaching CLIL classes are non-native speakers of the target language who are professional specialists in the content areas they teach in CLIL classes (Dalton-Puffer, 2011); content-wise, this is a merit since it facilitates student initiations into target discourse communities. Coyle (1999) has argued that CLIL is founded on four pillars: (1) Content, (2) Communication, (3) Cognition, and (4) Culture (i.e., the CCCC or 4C's model of education); it aspires after (a) teaching and learning subject matters, (b) learning and using language, (c) developing learning and thinking skills, and (d) fostering social awareness of self and others. Coyle's 4C's model of education shows the broad definition of CLIL (which is cognitively demanding for teachers). In fact, Coyle proposed this model as an aid to teachers who plan to start teaching CLIL classes. A shortcoming of such a broad definition is that teachers find it difficult to distinguish between CLIL and 'good teaching' practices. While CLIL requires the simultaneous teaching of both language and specific subject matter, 'good teaching' practice is a question of 'how well' these are taught. As such, good teaching practice in a CLIL classroom requires that the teacher be engaged in reflective teaching, and that the learners be engaged in effective learning (cf., Salmani Nodoushan, 2006). It is the teacher's job to (a) teach specific subject matters, (b) teach language, and (c) make sure students learn and understand both.
It should also be noted here that CLIL is not a single method, but that it covers a range of methods of, and approaches to, effective FL teaching that vary in terms of complexity. The simplest form of CLIL, called the 'language bath', claims that language is learnt if students are immersed in the target language in CLIL classrooms (Dalton-Puffer, 2007). By way of contrast, more elaborate approaches to CLIL hold that simple immersion is not adequate, and that students must be engaged in more cognitively-demanding activities and tasks. Lyster (2007), for instance, argued that accurate language learning is not fostered in students unless the teacher engages in 'focus-on-form' activities and provides corrective feedback where needed (cf., Gholami, 2017;Khalili & Mohammadnia, 2019;Salmani Nodoushan, 2010). Likewise, Swain (1995), proposed the 'output hypothesis' which assumes that students will not learn a language unless they are urged to produce spoken and written output in that language. From a theoretical perspective, the 'penta-pie model' (Westhoff, 2004) is even more complex. According to this model, effective foreign language pedagogy is not possible unless it (1) exposes students to meaningful input, (2) engages them in form-focused processing, (3) involves them in meaning-focused processing, (4) provides them with opportunities for output production, and (5) makes them aware of, and trains them to use, language learning strategies. De Graaff et al. (2007) brought this model to bear on language teaching in a CLIL classroom and argued that all five aspects should work in tandem to ensure optimal learning.
In recent years, CLIL has witnessed a tendency towards the teaching of disciplinary literacies-i.e., discipline-specific discourse and genres (Meyer et al., 2015;Morton, 2010). This is where ESP and CLIL merge in terms of methodology. Nevertheless, immersing students in an 'ESP tub' does not guarantee their mastery of disciplinary literacies; such literacies need to be explicitly taught in the CLIL classroom, and teachers should get their students engaged in tasks and hands-on experiences that guarantee optimal language and content learning. Only then can CLIL create an innovative learning environment conducive to the simultaneous acquisition of both a language different from the students' mother tongues and subject matter knowledge (cf., Francomacaro, 2019;Martín Del Pozo, 2017;Naddeo, 2019). As Johns and Salmani Nodoushan (2015) rightly put it, such an approach would respond to learner needs in the 21st century provided that CLIL employs effective, on-going, needs assessment and target situation analysis (cf., Anthony, 2018;Mobashshernia, 2020).
As Kampen et al. (2018) rightly said, "CLIL pedagogy is typically associated with innovative, student-centered pedagogical approaches. Focusing on studentcenteredness is considered necessary for effective subject and foreign language learning to occur" (pp. 223-224). It was stated earlier that ESP was learner-centered in the past, but that it has more recently become 'learning-centered'. Proponents of the integration of CLIL into ESP programs need to clarify which perspective they would opt for: (a) the student-centered, (b) the learning-centered, or (c) a combination of both. It should also be noted that, in spite of their apparent similarities in ESP contexts, CBI and CLIL differ in that the former is so devoted to the 'content' itself that it fails to heed (a) the values and/or genres of content-creators, and (b) the strategies that learners need to master for their successful exploitation of the content area (Johns & Salmani Nodoushan, 2015); nevertheless, there is much more to ESP than mere content. Casanave (2017) and Tardy (2016)-among others-have argued that CBI creates problems at the stage of student initiations into their respective discourse communities (cf., Johns & Salmani Nodoushan, 2015).

Genre-based teaching in ESP
Genre-based teaching was in essence a response to process approach to teaching writing by White and Ardnt (1991). According to White and Ardnt (1991), writing is a process that comprises six interconnected stages: (1) focusing on the real reasons for writing; (2) structuring the organization of ideas with an eye on the target readership; (3) drafting, or the transition from writer-based thought into reader-based text; (4) reviewing, or standing back from the text and looking at it from the perspective of the target readership; (5) evaluating, or the assistance the teacher provides for the students during the writing task; and (6) generating ideas in the form of written text (White & Ardnt, 1991).
Proponents of the genre-based approach to teaching argued that process approach by White and Ardnt (1991) is short-sighted in that it fails to address (a) the unique requirements of specific writing tasks, and (b) variations in individual writing situations (Paltridge, 2013). By way of contrast, opponents of the genre-based approach argued that it is too much product-based (i.e., product-oriented) and assumes fixed patterns for texts; they criticized the genre-based approach on the assumption that it motivates learners to look for pre-determined formulae and fixed patterns which they can then bring to bear on their task performance.
In an attempt to save genre-based teaching from such criticism, Flowerdew (1993) tried to dovetail process approach and genre-based teaching; he suggested that the fixed, rule-governed patterns of specific genres should be considered as prototypes that can be used with variation in the process of text production. He further suggested that genre-based teaching is concerned with 'learning about' genres, not their end products (Paltridge, 2013). Likewise, Badger and White (2000) argued that genrebased and process approaches should work in tandem for best results since they are complementary, but not complimentary. Johns (2008) suggested that the genre-based approach has two jobs: (1) to focus on and promote genre awareness, and (2) to culminate in genre acquisition. The former will ensure that students will acquire the skills and strategies that they need to be able to respond to varied novel situations and tasks; the latter will ensure that they acquire the professional competence which they need for task performance. Likewise, Hammond and Mackin-Horarick (1999) have emphasized that genre-based teaching paves the way for students to access the skills, strategies, professional knowledge, texts, and discourses that are crucial for their successful participation and acceptance in their target discourse communities.
Nevertheless, there are others who warn us about the potential threats of teaching 'genres of power' through the genre-based approach. Genres of power, in professional communities, can be defined as discursively-recognized linguistics conducts (e.g., texts, speeches, etc.) of which the job is to create an imbalance in authority and/or power. The term 'genres of power' has been borrowed from sociology where it has been defined, à la Kratz (1989), as culturally-recognized ways of linguistic behavior (i.e., speaking or writing) of which the constitutive pragmatic definition inevitably includes a difference of power or authority (e.g., only elders have the right to bless). By this token, Luke (1996) has implicitly warned us that teaching the fixed, rulegoverned patterns of a given genre in a genre-based classroom may potentially culminate in the uncritical and mechanical reproduction of its status quo (cf., Paltridge, 2013). Belcher (2006), however, has suggested that the explicit teaching of genres (a) gives learners access to the resources that they need for professional development, and (b) prepares them for professional participation in the work, study and everyday-life worlds of their imagined target discourse communities.

WHERE ARE WE NOW? JIT VERSUS JIC
Now that we have a clear picture of the theoretical, analytical and methodological developments of ESP in the past few decades, we can ask a fundamental question: Where are we now? Pennarola (2019) argued that the current state and the future direction of ESP will be defined by epistemological attempts that seek to bridge the gap between the academy and the workplace. Likewise, Johns and Price (2014) suggested that ESP, as it is practiced around the world today, comprises English for Occupational Purposes (EOP) and English for Academic Purposes (EAP), both of which are similar in that they both require on-going needs assessment and target situation analysis (Johns & Salmani Nodoushan, 2015). However, the main difference between the two lies in their perspectives on teaching ESP. While EOP is directed at "just in time" (JIT) learning (i.e., learning through trying), EAP preaches "just in case" (JIC) learning (i.e., learning through training).
Just in case learning prepares students for a profession in some unknown future, but just in time learning is learning through 'hands-on experience' which takes place once the learner has already entered the target profession. Just in time learning is worthy in that it (a) is highly motiving, and (b) prepares learners for the immediate application of what they learn through hands-on trials or in professional development workshops. However, not all of the skills and competencies that a profession requires can be learnt just in time. In many cases, the skills, competencies and patterns that a profession calls for will have to be learnt just in case so that the learner has enough time to subsume and internalize them and then convert them into procedural knowledge (Brandenburg & Ellinger, 2003;Govindasamy, 2001). Proponents of just in time learning argue that this approach allows the learner to learn through discovery; once learners feel the need to learn, they start the discovery process and learn what they need. This process is conducive to deeper learning because learning takes place through a problem-solving process where joy, motivation, involvement and concentration work in tandem. JIT allows learners to become deeply absorbed in the learning process since they feel the pressing need to solve a real problem, perform a meaningful task, or accept a real challenge that matters to them (Govindasamy, 2001).
While I do agree with the proponents of JIT learning about its merits in relation to certain tasks, I do not see it realistic to assume that language can be learnt at exactly the time when it is needed. Language learning is a painstaking and time-consuming gradual process that requires a lot of devotion, work, and practice. To me it seems that JIT has entered educational settings as a result of the 'commodification' impact of capitalist economic systems which have transformed goods, services, ideas, and people into commodities or objects of trade.

FUTURE DIRECTIONS
ESP has come a long way so far, and it still has a much longer way to go in future. Like its past, ESP will develop and proceed in theoretical, analytical, and methodological pathways in the years to come.
From a theoretical perspective, ESP will opt for an epistemological perspective which requires the ESP researcher to be a surveyor whose job is to map specialized subject domains (Détourbe, 2017). Such researchers will adopt a wide-angled approach on ESP research and will embark on a systematic and contextualized survey of the (a) discursive, (b) generic, (c) social, and (d) organizational dimensions and features of the specialized domain which they put under scrutiny. They will aim at building a comprehensive picture of the professional and specialized varieties of English that are used within that domain. To this end, they will borrow operative concepts and methodological tools from other disciplines and integrate them to shed light on the specialized domain they seek to study (Détourbe, 2017).
From an analytical perspective, ESP will continue to be informed by corpus linguistics/analysis, discourse analysis, and computer and digital technology. The importance of fieldwork and contextualization will be emphasized in ESP research, and genre analysis will most probably be combined with an ethnography-driven approach (cf., Domenec, 2017). Some future ESP studies will focus on the (a) milieu, (b) discourse, and (c) culture of the discourse communities-that will be scrutinized. Since policy, peace, health, services, information, and the environment are important topics in the post-industrial society (Salmani Nodoushan, 2020b), ESP research is expected to be responsive to these topics, and it will not be surprising if some ESP researchers decide to put such topics under analytical scrutiny. The discourse of the corporate world might be another area for ESP research, and the context that shapes corporate-world culture and communication may be researched (cf., Domenec, 2017). Such studies may use discourse analysis complemented by other approaches as an entry point into the survey and analysis of specialized communities. Needs analysis and target language use situation analysis will also continue to inform and shape ESP research and materials in the years to come.
From a methodological perspective, ESP may give more attention to the popularization of specialized knowledge. This has already been under way for some years-e.g., Technology, Entertainment, and Design (TED) talks. Specialized texts that will be developed with the aim of popularizing specialized knowledge will simplify complicated specialized content and present it in simple language for a public audience. This may culminate in the simplification of academic language and will also impart academic literacy to the wider public audience (cf., Mattiello, 2017). It is predictable that attempts at the popularization of specialized knowledge will have to (a) reduce the specificity of scientific language or jargon and opt for less technicality in content and vocabulary, (b) avoid the serious tone of specialized texts and allow for a conversational, humorous and less formal or informal tone, and (c) opt for a narrative style in texts instead of the current informative, expository, or argumentative styles (cf., Mattiello, 2017). E-learning technology will also affect ESP, and just in time learning will become ever more prominent.
Nevertheless, ESP will continue to be available in different forms; the main types of ESP that are likely to exist in the years to come, à la Détourbe (2017), are: English for Academic Purposes (EAP), English for Specific and Academic Purposes (ESAP), English for Research Publication Purposes (ERPP), English as an Academic Lingua Franca (EALF), English for Academic and Occupational Purposes (EAOP), and English Medium Instruction (EMI).
Last but not least, ESP will continue to be based on four pillars already delineated by (Anthony, 2018; see also Mobashshernia, 2020): (1) Needs analysis: posing questions about the real wants, lacks, and necessities of target learners; posing questions about the environment and context in which learning is expected to take place; (2) Learning objectives: determining the language, genres, and skills that target learners might be expected to master and use in a target situation; developing sets of learning strategy objectives for the ESP class; (3) Materials and methods: developing a teaching methodology, specialized textbooks or other types of materials (e.g., audios, videos, etc.); (4) Evaluation: measuring learners' performance gains on a test; evaluating the success of the ESP course and the quality of instruction.

CONCLUSION
Based on what went before, it can be concluded that genre analysis has had, and will continue to have, a pivotal role in shaping the status quo of ESP. Swales and Feak (2012) have specifically done a great job for graduate students, but the need for more sophisticated genre studies is still felt in all areas and also for all students (cf., Johns & Salmani Nodoushan, 2015). All in all, it can be suggested that, in the years to come, ESP specialists need to be open, flexible, and sensitive to context. They also need to be very good learners.