Discourse Studies and Urban Research: Methodological Challenges, Achievements, and Future Prospects

: This paper offers a critical appraisal of the methodological capacity of Discourse Studies (DS) in conducting urban research. Based on an extensive literature search, 125 publications that explicitly claim to utilise DS were reviewed. The results show that DS has been utilised for its methodological value, critical lens, interdisciplinary approach, ability to reveal the undiscovered, and presentation of new insights to urban questions. This paper identiﬁes and discusses major sources of inspiration and main trends in utilising DS in urban research. Theoretical diversity, the scarcity of analytical framework, and the lack of required expertise and skills are presented as three main methodological challenges for urban researchers. This paper concludes with suggestions for advancing the use of DS in urban research: obtaining an in-depth knowledge about its theoretical foundations, gaining an analytical overview of the strengths and weaknesses of different approaches, developing innovative frameworks that better explain urban questions, and gaining required linguistic knowledge for the application of DS.


Introduction
There has been a growing interest among researchers in different areas of urban studies to explore the significance of language and discourse in shaping, debating, understanding, and changing urban policies. This growing interest, which reflects the wider desire of urban scholars to benefit from qualitative methods, was a response to the dominant technical and technocratic approach to planning and advocated a shift towards the recognition of the critical role of language and discourse in human affairs [1]. As the result of this "linguistic turn" in the social sciences [2], language lost its neutral status, promoted by the positivist tradition, and was problematized "as a medium, a system of signification through which actors not simply describe but create the world" [3] (original emphasis). In the field of urban studies, the role of discourse as a component of urban processes and urban change was highlighted [4], and researchers have tried to integrate the study of language into urban research and analysis [5].
Advocates of this approach to urban studies believed that "language has the capacity to make politics, to create signs and symbols that shift power balances, to render events harmless or, on the contrary, to create political conflict" [6]. They argued that policies are developed, debated, negotiated, and discussed "in" language, and communicated and implemented "through" texts, and this underlines the significance of the discursive approach to understand and analyse policy making. As Weiss and Wodak [7] put it, "texts are often sites of struggle in that they show traces of differing discourses and ideologies contending and struggling for dominance". Policy discourse, thus, is constructed by a bundle of exchanges that shape policy-making processes and debates through metaphors and practices [8]. On the other hand, deliberate changes in the language of politics-making brings about changes in the broader socio-political reality. Textual changes, in the form of metaphors, terminologies, and new statements, challenge existing realities and call for transformation: "Decisions to adopt new textual modes and styles are also decisions to alter the discursive-and therefore the broader-practices which such texts realize." [9].
In the early 1990s, a number of scholars established a new branch in discourse studies that is referred to as critical discourse analysis (CDA) [10]. CDA explores the dynamic relationships between discourse and social power [11] and is interested in exploring the manifestation of dominance, discrimination, power, and control in language [7]. According to Fairclough, what unites these scholars "is a shared interest in the semiotic dimensions of power, injustice, abuse, and political-economic or cultural change in society" [12]. CDA has covered a wide range of issues, including racism, immigration, neoliberalism, globalisation, gender, education, doctor-patient communication, war and terrorism, and welfare and unemployment, to mention but a few [13].
Despite the growing number of publications in urban research that utilise discourse studies (DS) as the departure point and the interest of researchers to explore urban issues from the perspective of discourse studies, the methodological merits of using DS in this field have not been analysed. Although these publications provide evidence for the centrality of language in urban processes on the one hand and the significance of DS in exploring urban questions on the other, there is no analytical study that critically reviews this body of literature. A timely review of the methodological contribution of DS to urban research will provide researchers with signposts for the previous works of scholars, offer guidance for scholars who want to deploy DS in the future, and provide a pathway for future research in this field. Thus, to address this lacuna of knowledge, this paper asks two critical questions: why and how DS has been utilised in urban research. While the former investigates the methodological capacity of DS in conducting urban research and its potential in responding to different urban questions, the latter looks at different ways in which DS has been utilised and different methodological frameworks that have been employed or developed. This paper first provides a brief overview of DS, followed by an explanation of the research methodology. Then, the findings of this study are presented in order to explore why DS has been utilised, how different methodological frameworks have been employed or developed, and which approaches to DS have been the main sources of inspiration and why. The concluding section underlines the main methodological challenges of using DS in urban research and suggests how these challenges can be addressed.

Discourse Studies, A Brief Overview
As noted, studying the role of language in urban studies has encouraged researchers to benefit from the rich and rapidly developing body of knowledge produced under the umbrella of discourse studies (DS). It is beyond the scope of this article to review the history of DS, different schools, methods, etc. However, some hints to the key elements will be helpful.
The term "discourse analysis" (DA), or "discourse studies" (DS), as we use in this paper following Van Dijk [14], has been used in a range of social science disciplines to study and analyse a variety of elements related to communication, language use, social relationships, etc. Scholars have suggested different definitions for DS, which is due to a plethora of understandings about the term "discourse" [15]. To give some examples, DS is about talking and text in context [16], refers to the study of language in the everyday sense that people use it [17], and investigates the relationship between language and the social and cultural contexts in which it is used [18]. Despite the fact that language plays a pivotal role in analysing discourse, we should not delimit "discourse studies" to "language analysis". Discourse is more than letters and words and combines elements of language with other elements of semiosis, including gestures, eye gaze, rhythm and tone of speech, and fluctuations in voice [19]. This is what scholars discuss as "multimodality of discourse". According to Kress and van Leeuwen [20], multimodality refers to the "use of several semiotic modes in the design of a semiotic product or event, together with the particular way in which these modes are combined".
Broadly speaking, there are two strands of understanding discourse: discourse as linguistic utterance and discourse as knowledge [21]. The first strand is more concerned with textually oriented discourse analysis [22], in the sense that it analyses all sorts of written and spoken texts. Linguistic traditions have seen discourse as the units of written and spoken communication and focused on the content of texts and conversations. In the second strand, which draws on post-structural theory and particularly the works of Michel Foucault, discourse refers to different ways of structuring areas of knowledge and social practice. As Keller suggests, a Foucauldian perspective on discourse is more interested in knowledge than in language [23]. In Foucault's work, the idea of discourse gains changing meanings. In his early "archaeological" work, discourse constitutes and constructs various dimensions of society, such as the objects of knowledge and conceptual frameworks. In his "genealogical" works, discourse is political as the power struggle occurs over and in discourse. According to Foucault, "power is exercised from innumerable points" [24] and "each society has its regime of truth, its 'general politics' of truth: that is, the types of discourse which it accepts and makes function as true and false statements, the means by which each is sanctioned; the techniques and procedures accorded value in the acquisition of truth; the status of those who are charged with saying what counts as true" [25].

Methodology
A literature search was conducted using keywords "urban" and "discourse" in two electronic databases, namely Google Scholar and Science Direct. To be sure that all relevant publications were included in the list of the potential literature for analysis, keywords were extended to cover phrases such as urban discourse, urban studies, urban planning, city, etc. This exercise generated a list of 189 items. Since the aim was exploring the methodological contribution of DS to urban research, the next step was identifying publications that explicitly claimed to utilise DS as the analytical tool. A total of 125 publications (one hundred and twenty peer-reviewed articles, four book chapters, and one book) were eventually selected for more in-depth analysis. No time span was set in order to find all the publications within this area. However, all the selected publications are from 1990s onwards, with the majority of them (104 out of 125) published after 2000. This trend could partially be due to the digitalisation of publications and hence their availability. However, this could also indicate the growing attention of urban scholars regarding the use of DS in urban research in the last two decades. Two aspects were the focus of the analysis: why authors utilised DS in their research and how DS was utilised. All the selected publications were carefully reviewed and analysed based on these two themes. There are some limitations for this analysis. First, it only includes publications in the English language. Second, there may be publications that have not been included in this research. Expanding the search could potentially lead to a more comprehensive list of analysed publications. However, after using two different search engines and expanding relevant keywords, this analysis reached a point of saturation at which no new relevant articles came to light.

Results and Discussion
This section discusses findings regarding the proposed key questions. First, we will explore the methodological capacity of DS in conducting urban research and discuss why DS has been utilised. Then, we will show how different methodological frameworks, informed by DS, have been employed or developed. Finally, we will show which approach to DS, or which methodological frameworks, have been the key sources of inspiration for scholars and what the strengths and weaknesses of these frameworks are.

Why Has DS Been Utilised?
Our analysis suggests that different researchers have utilised DS in their research for different reasons and purposes. However, the most common reason has been to use it as a method: a "methodological tool" [26], a "critical tool" [27], a "supplementary method" [28], "a useful set of methodological tools" [29], "a vehicle of policy and project analysis" [30], a "theoretical and methodological perspective" [31], and an "alternative" method [32]. To utilise DS as a method of inquiry, researchers refer to the varied potentialities and capacities of DS. Some scholars find the "critical" dimension of DS relevant to their study. For example, in the field of housing research, the critical potential of DS enables researchers to investigate emerging issues of power and resistance [27]. In the field of policy making, it problematizes the linguistic, identity, and knowledge dimension of policy-making practices [33]. Others refer to the "relational" capacity of DS, in that it links together two subject areas and offers new insights. For example, DS provides space to discuss how material and discursive changes are related at multiple levels of analysis [34] and explores "dialectical relationship between discursive and social practices of the actors involved in pluralistic forms of governance" [26].
In some studies, scholars have underlined the strength of DS in unearthing and revealing the hidden and the undiscovered. For example, DS can highlight "silence in the text" through uncovering those voices that are (mis)represented [35]. It can examine hidden dimensions and unintended consequences of social actions [28] and "uncover the power and economic realities that belie major developments and policy directions" [29]. DS allows us to discern implicit assumptions about the current state of affairs and the future of the world [36] and helps researchers to reveal how different policy actors and agencies construct and interpret various meanings to frame priorities and proposed policies [37]. Moreover, through the analysis of language, DS enables the researcher to unfold representations and hidden mechanisms that create specific narratives through analysis of the language [38] and unmask hidden ideological agendas that generate political text and talk [39].
A number of scholars have highlighted the potential of DS in opening up new insights and perspectives to the subject study and the discipline. For example, it has been argued that DS identifies new sites of politics and is able to analyse their political dynamics [6]. Moreover, DS provides new insights about the relational and constitutive aspects of policy making and the interaction mechanism between organisational structures and agents [28] and increases awareness of practitioners in the planning practice regarding the power of discourses produced by both planning authorities and developers [29]. DS has the potential "to go beyond academic analyses of texts by highlighting 'bottom up' practices of resistance" [40] and can assist "policy-makers in reconceptualising their approach to problems and, most importantly, to understand why certain issues come to be perceived as 'problems'" [28]. It can also provide new insights into place identity by taking cultural mapping to a deeper level and incorporating impressions and stories into the process of building place identities [31]. Table 1 presents an overview of the rationale behind utilising DS in urban research. As this table suggests, and taking into account our discussion above, one can argue that DS has been attractive for researchers for following five main reasons: its methodological capacity offers them an appropriate conceptual framework, its critical dimension provides them with a fresh critical lens, its relational aspect constructs an interdisciplinary basis for research, it provides them the power to uncover the undiscovered dimensions of a problem, and it grants new insights and perspectives to the study of an urban question. However, as will be discussed in detail in the next section, approaching DS as a method has not been straightforward and has its own challenges and dynamics.

Purpose Examples
Methodological capacity "I suggest that critical discourse analysis can provide the methodological tools to explore the dialectical relationship between discursive and social practices of the actors involved in pluralistic forms of governance" [26] "Discourse analysis provides the social scientist with a supplementary method that can reveal the different ways that terms and concepts are drawn upon to justify urban policy intervention" [28].
"Techniques of CDA provide a useful set of methodological tools for analysing planning documents to uncover the power and economic realities that belie major developments and policy directions" [29].
"The researchers chose DA as a vehicle of policy and project analysis" [30].
"[W]e understand DA as a more theoretical and methodological perspective, than a rigid set of methods" [31].
"The findings of this study suggest the use of a discourse analysis approach as an alternative to a thematic one in interpreting residents' views of the gentrification process in their neighborhoods" [32].
iscourse analysis does have critical potential and, used appropriately, offers opportunities for critical housing researchers to explore new issues of power and resistance in the housing field" [27].
" [D]iscourse analysis problematizes what conventional policy analysts take for granted: the linguistic, identity and knowledge base of policy making" [33].
Relational capacity "Critical discourse analysis draws attention to the constitutive effects of language on constructing welfare subjectivities, as well as providing a space to discuss the relationship between material and discursive change at multiple levels of analysis" [34].
"I suggest that critical discourse analysis can provide the methodological tools to explore the dialectical relationship between discursive and social practices of the actors involved in pluralistic forms of governance" [26] Strength in revealing the undiscovered "A critical discourse analysis has been able to highlight 'silence in the text'-the voices of those that are (mis)represented" [35].
"Discourse analysis entails an obligation to examine not just what is apparent at a superficial level but also the hidden and unintended consequences of social action" [28].
"Techniques of CDA provide a useful set of methodological tools for analysing planning documents to uncover the power and economic realities that belie major developments and policy directions" [29].
"CDA allows the analyst to discern implicit assumptions, about how the world is or should be" [36].
"In relation to nature and urban planning, discourse analysis potentially enables the research to 'reveal' how policy actors and agencies have constructed and interpreted various meanings of nature and therefore 'frame' both analysis of nature or environmental and urban problems, the identification of priorities and proposed policies" (Duvall et al., 2018, pp. 490-491) [36,37].
Discourse analysis "allows to unfold representations and hidden mechanisms behind creation of specific narratives through analysis of the language, particularly the use of vocabulary, rhetoric, communication conventions and patterns which are then described, interrelated and explained" [38].
"These 'critical discourse analysis' approaches are motivated by an ambition to unmask hidden (e.g., capitalist, right-wing) ideological agendas as drivers of political text and talk, to advance democratic stakeholder participation in decision making and to critically analyze discriminatory (e.g., racist, antisemitic) language use, especially in the public sphere or by political actors" [39].
Opening new sights and perspectives "Discourse analysts are well equipped to identify the new sites of politics and analyse the political dynamics therein" [6].
"Discourse analysis provides a method of exploring how policy conflicts are routinised, negotiated and acted on. Such an approach can help provide new insights about the relational and constitutive aspects of policy-making and the ways in which organisational structures and agents interact" [28].
"The application of CDA to spatial planning policy can extend planning practice, making practitioners more aware of the power of discourses, both those produced and owned by the planning authorities, and those authored by developers" [29].
"Critical discourse analysis (CDA) is an approach to discourse that is particularly suited to advocacy based research because it has the potential to go beyond academic analyses of texts by highlighting 'bottom up' practices of resistance" [40].
"A discourse analysis can assist policy-makers in reconceptualising their approach to problems and, most importantly, to understand why certain issues come to be perceived as 'problems' [28].
"Discourse analytic thinking . . . emphasizes time and space as they are imagined, not measurable, concepts, and lets impressions and stories come to light as a legitimate part of building place identities" [31].

How Has DS Been Utilised?
Overall, we can identify four trends in utilising DS as research methodology: (1) studies that claim to borrow theoretical frameworks and analytical techniques from the exiting literature of DS but fail to illustrate how they have been utilised; (2) studies that explicitly apply existing methodological frameworks and techniques to their research and build their argumentations and discussions on them; (3) studies that are inspired by existing analytical frameworks but generate their own that draws on the theoretical and technical knowledge of DS; and (4) studies that criticize the existing methodological frameworks of DS in their current forms and claim to establish an alternative or modified framework.
In many cases, researchers claim to utilise DS as a methodology but fail to introduce a clear methodological framework. In these cases, it is not clear how discourse is analysed, what are the techniques used for the analysis, and what the key themes of analysis are. Several studies fall into this category. Franklin's project of using language and discourse analysis to examine different perspectives on definitions of housing quality and good housing design in the UK fails to suggest a clear definition of discourse or a methodology to analyse policy documents [41]. Matthews and Satsangi [29] do not present a framework for analysing texts, documents, and marketing materials produced regarding the Leith Docks redevelopment project and do not clarify which techniques of text analysis are employed.
Winkler [42] claims to use CDA to investigate the apparent assumptions underpinning the City of Johannesburg's planning policies in order to demonstrate who economically and spatially benefits from public sector-led regeneration programs. However, there is no reference to the existing literature on CDA, nor is a framework introduced to show how the analysis is conducted. Xu [43] claims to contribute to the lack of scale dimension in discourse analysis and suggests scaled discourse analysis to address this lacuna. However, there is no dialogue regarding the wider literature of DS, nor is a methodological framework proposed. In their study, Bunders and Varró [44] claim to combine Foucauldian discourse analysis and Laclau and Mouffe's discourse theory, but no clear methodological framework is introduced to show how these approaches have been combined and what exactly the methods and techniques are.
The second trend of research includes studies that explicitly borrow methodological frameworks from prominent researchers of DS and apply them to their study. For example, Darcy [45] investigates manifestation and reproduction of three concepts of "democratisation", "commodification", and "technologization" in micro-level discursive and social practices in the policy documents related to the development of community housing policy in Australia. This framework draws on Fairclough's [22] discussion on broad tendencies in discursive change affecting the social order of discourse, which reflects the general directions of social and cultural change. Lacerda [46] analyses the discourse on favelas produced by Brazilian society and consumed in the political field of local administration. To achieve this, three specific questions are raised, and for each, a clear methodology is suggested using Wodak's discourse-historical approach [47] and Fairclough's dialectical-relational approach [48]. A good example of research that establishes a robust and clear methodological framework using the DS literature and applies it to the discussion and analysis is the case study research conducted by Marston [35,49], who investigates the changing nature of public housing policy in Queensland, Australia, and explores the discursive constructions of the policy problems and power relations within the policy community. Marston uses Fairclough's [22] tripartite structure, which argues that any instance of discourse is simultaneously a piece of text, an instance of discursive practice, and an instance of social practice. Textual analysis is concerned with the form and meaning of text, discursive practice focuses on the discursive production and interpretation of text, and sociocultural practice operates at the level of broader social analysis.
In the third trend, researchers develop a holistic, methodically structured analytical framework, substantially informed by the DS literature. In this way, the proposed analytical framework is not a direct employment of an existing analytical framework but draws on the knowledge and techniques of DS and develops a new one. In his study, Collins [50] aims to analyse the changes that took place between the government and the stewards of the Upper Clyde Shipbuilders in 1971. He draws on Voloshinov's notion of "evaluative accent"-the idea that every act of utterance involves the value judgement of the speaker, which is inherently social in nature. A well-structured methodology is presented by Klodawsky and colleagues [51] in their study of four years (1994 to 1997) of media reporting on homelessness in The Ottawa Citizen daily newspaper. Informed by Callahan and Callahan's [52] interpretation of van Dijk's [53] work, the study conducts a textual analysis of the reports using five lenses of macrostructural analysis, intertextual analysis, grammatical analysis, relevance structuring, and rhetorical structures [51]. Adscheid and Schmitt [54] analyse two Swedish key projects of sustainable urban development in order to explore the role of the actors involved and the post-political character of the environments envisioned by these actors and to investigate how these environments are mobilized and legitimized across spatial and institutional contexts. To achieve these goals, they link sociology of knowledge approach to discourse analysis (SKAD) with actor network theory (ANT). In their study, the four-step discourse analysis proposed by Keller [23] provides insights into the infrastructures of discursive knowledge production and problem solving within the case study projects and explains how the institutions that are part of the actor network sustain and expand these infrastructures.
In the fourth trend, scholars have a critical approach to DS and claim to develop alternative frameworks. For example, Goodchild and Cole [55] examine the meaning of social balance in social housing practice in Britain and argue that concepts of social balance are multi-layered and cannot be properly treated as the mere product of political ideology in the way suggested by DS. They propose developing a modified form of post-structural discourse analysis that suggests truth is relative. The proposed framework draws on three levels of social reality suggested by Lefebvre and intends to analyse social balance on three levels of national policy, estate management and upgrading, and the social experience of residents. Another example is the work of Rydin [56], which proposes a rhetorical methodology of policy discourse analysis in order to examine the locus of power in the relationships between central and local government and the key economic interests within the Thames Gateway area of London. The methodology draws on rational choice approach presented by Dryzek [57] and Dowding et al. [58], which understands communication as an actual or implied argument and intends to reveals the active role of language within the policy processes.
It needs to be noted that computer-aided programs have been increasingly employed as a tool for comprehensive analysis and data processing. For example, to investigate the importance of communicative and participatory paradigms in contemporary European territorial policy, Damurski and Oleksy [59] analysed 10 policy documents benefiting from natural language processing (NLP) tools implemented within the CLARIN infrastructure, a software that helps to understand analysed texts. They also utilised the web-based system Inforex for the configuration of the corpus and setting the annotation environment. In some cases, data analysis software and programs have been used to facilitate the data organizing, coding, and content analysis of texts and documents, such as NVivo software [37], ATLAS-ti software [60], AntConc tool [61], and MAXQDA software [38].

Key Sources of Inspiration
Our analysis found that some key figures and conceptual frameworks have been intensively utilised by scholars and have made a significant contribution to the advancement of DS in urban research (See Table 2). The three-dimensional model proposed by Fairclough [22] has been one of the frameworks most utilised by researchers. According to Fairclough, discourse analysis should explore the relationships between three dimensions: text (different types of written, spoken, and visual materials), discursive practice (various aspects of the process of text production, distribution, and consumption), and sociocultural practice (including the immediate situational context of a communicative event, the wider context of institutional practices the event in imbedded within, or even wider frame of the society and the culture). Subsequently, Fairclough [62] introduces three stages for the analysis: description (of formal properties of the text), interpretation (of the relationship between text and interaction), and explanation (of the relationship between interaction and social context). Taylor [63] uses Fairclough's tripartite framework to analyse materials produced by public landlords to inform tenants about transfer proposals. This framework helps the researcher to explore discourse practice (macro level), text (micro level), and social practice (power relation) in a video, which was available for free on request to tenants whose house was under consideration for transfer by Scottish Homes. Another example is Arapoglou's [26] analysis of the dialectical relationship between power, discourse, and practice in the treatment of homelessness in Greece. Using official texts and a wide range of interviews with officials, NGOs, and some ethnographic interviews, Arapoglou introduces a clear methodological framework with strong argumentation and discussion. Davison [63] employs Fairclough's three-dimensional framework in his critical examination of approaches that are utilised by planning authorities in the city of Melbourne to regulate the "character" of property development. Hassanli et al. [64] apply Fairclough's tripartite structure to examine the representation of Airbnb in the local newspapers of Sydney communities with the largest share of Airbnb listings.
Several reasons can be named for the extensive use of Fairclough's approach to DS, particularly the tripartite structure he introduced. Firstly, the clarity of Fairclough's model makes DS comprehensible for a wide range of scholars. This clarity makes DS accessible for researchers, although linguistic techniques need to be learnt. As Marston puts it, "Fairclough's model goes some way towards making discourse analysis more accessible for a wider range of researchers, yet the craft of linguistic analysis must still be learnt" [35]. Secondly, the multidimensional nature of the framework that connects discourse to its production and socio-cultural context provides scholars with a comprehensive framework to explore urban questions that are inherently dynamic and multifaceted. Thirdly, the multiscalar character of the framework that starts with the text but connects it to the wider socio-cultural context helps researchers to explore the interconnectivity of the urban issues at different scales. Fourthly, the proposed analytical stages enable researchers to go beyond the confines of mere description and benefit from explanatory analysis and thus gain a better understanding from social phenomena and problems that are complex. Finally, this framework paves the way for an interdisciplinary analysis because it grants the researchers an opportunity to mix linguistics with other disciplines and benefit from their argumentation and research techniques.
In a number of studies, Hajer's notions of "discourse coalition" and "story line" have been utilised. According to Hajer, "A discourse coalition is thus the ensemble of a set of story lines, the actors that utter these storylines, and the practices that conform to these story lines, all organized around a discourse . . . Story lines are the medium through which actors try to impose their view of reality on others, suggest certain social positions and practices, and criticize alternative social arrangements" [3]. To give some examples, Lovell [65] examines how the framing of sustainable housing in the UK as "low-carbon housing" united around the two storylines of "life cycles" and "smart housing". Tozer and Klenk [66] analyse climate governance texts that were produced by 17 founding members of the Carbon Neutral Cities Alliance to identify the storylines underlying the urban imaginaries of carbon neutrality among these pioneers. To investigate how the shift from sustainability to smartness has reshaped urban strategies and interventions in the two cities of Malmö and Graz, Parks and Rohracher [67] use the two concepts of discourse coalition and storyline to understand existing sustainable city initiatives and investigate how the new discourse of smart cities changes the terms of the sustainability discourse and creates new openings and contested situations. Generally speaking, it is a usual practice that urban polices are constructed around some storylines and actors make coalitions to present one particular storyline as the official policy. Hajer's notions of storyline and discourse coalition provide urban researchers with useful tools to explore these storylines and analyse practices around discourse coalitions.
Another source of inspiration has been the discourse perspective presented by Robert Beauregard in his seminal work Voices of Decline [68], where he explores the post-war discourse of urban decline in the US and reflects upon the multiple discourses underpinning policy approaches to the cities. Beauregard's approach underlines the role of language (textbooks, political manifestos, political speeches, newspaper, and magazine articles, etc.) as an active element of constructing cities [69]. As Jacobs puts it, "For Beauregard, texts are instrumental in creating narratives of urban decline and his work has influenced a number of recent discursive approaches to the city" [70]. As an example, Boyle and Rogerson [69] examine the multiple ways in which the notions of power and discourse interweave with the production and legitimation of city development trajectories. Drawing on Beauregard's approach and framing of the post-war discourse of urban decline in the United States, they analyse city development strategies as a discourse and discuss its central characteristics, such as commodification and the sources of institutional power that both produce and reproduce the discourse.
Foucauldian approach to discourse has been also an inspiration source. Different reasons have been suggested for the usefulness of Foucault's approach to discourse for urban research. According to Jacobs, the Foucauldian approach "provides a firm basis for engaging in detailed archival scholarship and is perhaps best suited to more historical based analysis and for researchers who purport that discourse is a reflection of power relationships" [70]. Richardson argues that "Foucauldian discourse theory suggests that theories of planning have failed to adequately deal with power/knowledge dynamics in policy making" [71] (original emphasis). The Foucauldian approach to discourse allows a shift in emphasis from text to practices and actions [8] and is thus capable of understanding dynamics of power and knowledge in policy making and planning theory: "Foucauldian discourse theory offers a strong theoretical underpinning for discourse analysis methodology focused not on text, but on the social world" [71]. Flyvbjerg and Richardson [72] argue that the Foucauldian approach to power is relevant to planning theory, as it concentrates on "what is actually done" and helps us understand how space and power are closely bound up in planning: "The Foucauldian approach problematises existing planning tools and processes, suggesting the need for a power-sensitised understanding of the nature of knowledge, rationality, spatiality, and inclusivity in planning theory" [72]. In the field of public policy research, Hewitt [73] says that the strength of discourse analysis inspired by Foucault is the opening up of ways of understanding policy activity, which "emphasise the contingent nature of rationality and seek to uncover the power relations of policy making" [73].
As far as methodological framework for studying discourse is concerned, it can be said that the Foucauldian approach does not offer a straightforward and clear analytical framework. As Fairclough puts it, the changing meaning of "discourse" in Foucault's works, from "rules for constituting areas of knowledge" to "relationship between knowledge and power", makes it difficult to simply apply Foucauldian ideas to discourse analysis-his perspective should be put to work within traditional linguistically oriented discourse analysis [22]. In other words, Foucauldian concepts should be imbedded into frameworks presented by DS scholars. A good example is the work of Richardson and Jensen in a number of papers [74][75][76] where they set out a discourse analytical framework to explore how spatialities are "constructed" in spatial policy discourses in Europe. Drawing on Foucault and Hajer, they introduce three analytical spheres of language, practice, and power-rationality as the three dimensions of the analytical framework.

Contribution of DS to Urban Research: Methodological Challenges, Promises, and Future Prospects
In response to the question "Why has DS been utilised by different scholars in urban research?" our analysis found that the methodological capability of DS has been the main motivation. We also found that the key characteristics of DS make it a useful source of inspiration and a promising approach to address urban questions: DS offers a fresh critical lens, benefits from a relational aspect that provides an interdisciplinary basis for research, is capable of uncovering the undiscovered dimensions of a problem, and presents new insights in the study of urban questions. As DS has been used for different purposes by different scholars, we also see multiple ways in which principles of DS have been employed. We identified four main categories for the utilisation of DS in urban research. The first trend claims to borrow a theoretical framework from DS but is found to be the least successful approach: it fails to establish a strong dialogue with the DS literature and is unsuccessful in applying analytical frameworks borrowed from the CDS literature. Other trends make significant contributions to the employment of DS in urban studies. They either borrow existing analytical frameworks and implement them successfully or are informed by the CDS literature and propose a new framework. The large number of studies that fall into the first category, and thus provide only a weak link to DS and fail to employ and apply a clear methodology, indicate that utilising DS as a methodological tool, despite this being a main motivation for researchers to use DS, is not an easy task. In fact, both DS scholars and urban scholars have recognized methodological challenges and difficulties of DS.
DS scholars have argued that there is no theoretical orthodoxy in DS [13]. As Chouliaraki and Fairclough put it, it brings "a variety of theories into dialogue, especially social theories on the one hand and linguistics theories on the other" [77] and thus goes beyond methodologism and theoreticism. According to Wodak [78], DS is neither a homogeneous method, nor a school or a paradigm, but a shared perspective on performing discourse and linguistic analysis. It is more like a research program with various facets and theoretical and methodological approaches [79]. This theoretical and methodological diversity that exists in DS makes it challenging, particularly for researchers outside linguistic and discourse studies disciplines.
The methodological challenge of utilising DS in urban research has been acknowledged by urban scholars. For example, Lees [5] notes two interwoven challenges regarding the use of DS in urban research: ambiguous knowledge about the theoretical roots and methodological presuppositions of the claims about discourse and the lack of clear and detailed methodological structure utilised in the research. Jacobs [70] argues that some researchers are not clear about their conceptual suppositions and employed methods, and fail to back up their arguments with empirical evidence. The knowledge and linguistic skills required for its application have been also mentioned as a challenge for utilising DS in urban research [80,81]. Our analysis supports these arguments. As noted, a large number of the studies we analysed fall into the first category, as they fail to establish a strong dialogue with the CDS literature and remain unsuccessful in applying or developing a clear analytical framework. In fact, the penetration into the underlying message and rationale of a discourse necessitates using advanced DS techniques and a broad sociological knowledge, which is not within the expertise of many scholars educated in the field of urban studies. Thus, as Kumar and Pallathucheril [80] put it, few scholars have provided guidance about a specific method or introduced an explicit step-by-step methodology.
Overall, there are three key methodological challenges for utilising DS in urban research. First, the theoretical diversity of DS makes it difficult for urban researchers to gain easy access to the literature and attain a comprehensible and practical understanding from it. Second, because of the scarcity of clear and easily comprehensible and applicable methodological frameworks, it is hard for urban scholars to find the most appropriate framework from the large body of existing literature. Third, since DS needs specific knowledge about linguistics and advanced skills to use linguistic techniques, urban scholars generally lack such expertise and hence fail to properly employ existing frameworks and techniques. This is why, as our analysis suggests, in many cases, we observe neither an adequate dialogue with the DS literature, nor the employment of a clear methodological framework. However, the number of successful studies are promising-as was demonstrated, urban scholars have been able to either employ existing methodological frameworks or have been inspired by the DS literature develop their own frameworks and thus contribute to developing a rich and growing body of knowledge.
In Jacob's [28] concluding remarks of a Special Issue, he asks: "Is there a future for discourse analysis?" Our analysis and the growing number of studies provide us with persuasive evidence that DS has been a useful methodological reference for urban researchers and is able to offer a working framework for the analysis. However, there are still a number of major challenges, as listed above. To address these challenges, urban scholars should: obtain an in-depth knowledge about DS and its theoretical foundations; explore the diversity of DS and gain an analytical overview of the strengths and weaknesses of different approaches and analytical frameworks; benefit from the proposed DS models but explore the possibility of developing innovative frameworks that better explain urban questions; and finally, gain required knowledge (e.g., linguistic) and technical skills (e.g., computer-aided programs) for the application of DS. The future of DS in urban research depends on how scholars overcome these challenges. Data Availability Statement: Data sharing is not applicable to this article.

Conflicts of Interest:
The author declares no conflict of interest.