A Critical Perspective

Increasing numbers of researchers in the field of tourism studies are using discourse analysis as a means of critical investigation when faced with qualitative or textual forms of data, such as written documents, or visual materials such as photographs and brochures. Such data is often representative of how a group of people have made sense of and reflected on their own world and that of others. However, there are many different types of discourse analysis and many different ways in which discourse analysis has been deployed. This paper reviews the attempts of tourism scholars so far and argues that many researchers have utilized discourse analysis in an eclectic fashion. The paper thus differentiates between content analysis, simple textual analysis, semiotic analysis and post-structuralist forms of discourse analysis in order to provide a framework for other tourism researchers to use. In particular, it is argued that discourse analysis should proceed by recognizing that all texts are produced intertextually in relation to other texts, which are in tum embedded within power relations that give degrees of authority. Discourse analysis should thus treat texts as mediated cultural products which are part of wider systems of knowledge. It is argued that discourse analysis is not just interested in what is within the text itself but also in what has been left out and the 'secret' meanings that are not obvious. It is also argued that it is important to note the interruptions and disruptions that occur within the flow of qualitative data. Utilizing discourse analysis should mean the development of a more nuanced reading of the data and thus add a more critical edge to much tourism research.


Introduction
Ever-increasing numbers of tourism researchers are utilizing an array of discursive techniques with varying intentions, and varying effects. This paper provides a pathway through a broad interdisciplinary range of research methodologies that are often conveniently grouped together under the term discourse analysis. Tourism researchers have utilized forms of discourse analysis as a means of critically investigating representations of tourism experiences, destinations, motivations and practices. These representations include textual data and written documents, interview transcripts, photographs and brochures, all of which are illustrative of how a group of people have made sense of and reflected on their own world and that of others (Hannam 2003).
However, it is our contention that there are many different types of discourse analysis and many different ways in which discourse analysis can be deployed. Furthermore we argue that many tourism researchers have utilized discourse analysis in an eclectic and unsophisticated fashion. This paper thus differentiates between content analysis, simple textual analysis, semiotic analysis and poststructuralist forms of discourse analysis in order to provide a framework for other tourism researchers. It is concluded that discourse analysis should proceed by recognizing that all texts are produced intertextually in relation to other texts, which are in turn embedded within institutional power relations within, which degrees of authority are inherent. Discourse analysis should thus treat any textual or visual data as mediated cultural products which are part of wider systems of knowledge. Utilizing discourse analysis should mean the development of a more nuanced reading of textual data and thus add a more critical edge to much tourism research. The following sections of the paper will thus differentiate between forms of content analysis, textual analysis, semiotic analysis and discourse analysis for tourism research.

Content Analysis
Partly bridging the gap between quantitative and qualitative methodologies, content analysis is an empirical technique which involves the counting, identification of issues and interpretation of the content of a text which is assumed to be significant (Fiske 1990;Weber 1990). For example, content analysis could be used on an archive of tourist brochures to explore how different styles of people KEVIN HANNAM is Principal Lecturer and Head of Tourism at the School of Arts, Design, Media and Culture, University of Sunderland, Priestman Building, Green Terrace, Sunderland SRl 3PZ, UK. e-mail: kevin.hannam@sunderland.ac.uk DAN KNOX is Lecturer at the School of Arts, Design, Media and Culture, University of Sunderland, Priestman Building, Green Terrace, Sunderland SRl 3PZ, UK. e-mail: dan.knox@sunderland.ac.uk and places are represented at the same time or over time. Content analysis is a quantitative approach concerned with categorizing and counting occurrences of aspects of content. The basic assumption is that there is a relationship between the frequency of a specific theme and its significance or dominance. In contrast to post -structural discourse analysis, content analysis is a deductive research method that is utilized to find apparently pre-existing patterns within large volumes of material. Results tend to be presented using basic descriptive statistics such as frequency tables, charts and graphs for comparative purposes.
Doing content analysis will mean having to construct a coding frame. This is a set of themes into which material can be allocated. Possible categories could include the numbers, class, gender, and ethnicity of people represented in the material. Of course, the choice of themes depends upon the questions that the researcher wishes to answer. The construction of the coding frame is done by using a logical and consistent sample of the material, say ten per cent and may involve several re-workings until the researcher is satisfied that everything has been covered. The coding frame should be reliable in the sense that if two people were to analyze the same material with the same aims they would uncover the same results. The coding frame should also be exhaustive in the sense that all the data must be allocated to the themes, even if this means the use of an 'other' category. It is important to monitor what ends up in this category so that interesting results are not being neglected.
For example, in her content analysis of photographs in the Lonely Planet travel guide to India, Bhattacharyya (1997: 379-380) codes the photos into three groups of 'authentic sights', namely 'historical sites, natural world and social life'. She then re-codes the same material according to whether people are the focus of the photos, whether people are present but not significant, whether people are shown in a crowd, or whether people are absent. Her results indicate that over half the photos are of social life and that people are primarily the focus of this group of photos, in marked contrast to the other sights. This result should be the same if someone else analyzed the photos according to the same coding frame. Of course, different researchers are likely to generate different coding frames and thus different analytical outcomes.

Textual Analysis
Textual analysis is a qualitative technique concerned with unpacking the cultural meanings inherent in the material in question. The content of a text is seen as a mediator of latent and highly variable cultural themes. Textual analysis thus draws upon the researcher's own knowledge and beliefs as well as the symbolic meaning 24 systems that they share with others. The meanings inherent in any text are, though, always highly unstable. As Gregory and Walford (1989: 2) have argued 'texts are not mirrors which we hold up to the world, reflecting its shapes and structures immediately and without distortion. They are, instead, creatures of our own making, though their making is not entirely of our own choosing.' Compared with content analysis, textual analysis is often more time consuming and is usually conducted on a relatively small sample of material. The results are wordier and generally use verbatim quotations of the relevant material. Again, in comparison to content analysis, it is less standardized and is often seen as less rigorous, however, its strength lies in its detailed and in-depth analysis of specific cases. The validity and credibility of the research is maintained through the reflexivity of the researcher, i.e., the researcher needs to be constantly aware of the assumptions and preconceptions that are being made during the research and the possible impact of these on the results. Making the steps used as transparent as possible also helps in maintaining the integrity of the research and allows any conclusions to be better assessed.
Doing textual analysis usually means going slowly and thoroughly through the material a line or sentence at a time and attempting to think about what was meant and why. The material is then open-coded, i.e., as ideas emerge they are written down alongside the text itself. The aim is to get as close to the material as possible in order to avoid missing anything. As the themes accumulate they will hopefully spark off more theoretical ideas which can be followed up later. The notes are then formalized into categories or codes (usually abbreviations or colours). The segments will probably end up with several codes and can be of varying lengths. Some people cut up a copy of the materials and place the coded sections into piles. The idea is to organize the material so that interesting relationships or themes can be seen. Sometimes using simple matrices may help. This is an iterative process in that some of the codes will break down when it is found that a particular pile of material contains significant differences and needs to be recoded in more detail. For example, whilst researching heritage tourism Halewood and Hannam (200 1) developed an initial analytical framework for authenticity, but then recoded the textual data in order to tease out distinct types of authenticity.
Some qualitative computer programmes can help researchers to conduct both content and textual analysis by cutting, pasting and retrieving information i.e., they speed up the sorting of material and help to prevent the researcher from becoming snowed under with paper. While it is relatively easy to retrieve information using conventional word-processing programmes, some of the specific qualitative data analysis programmes, such as Nud.ist, also help researchers to create graphics or diagrams out of data so that the researcher can better visualize the relationships between codes or themes. However, computer programmes do not do the analysis, ultimately it is the researcher(s) who has to draw out the interconnections and explanations in the material in question. Using a computer though may speed things up and help researchers to manage their material more effectively (see Dey 1993;Kelle 1995;Fielding and Lee 1991;Weitzman and Miles 1995).

Semiotic Analysis
Tourism studies have, even noting the ocular-centric tendency of post -enlightenment thought, been particularly dominated by notions of the visual and representation. Fallowing U rry' s (1992) seminal account of the tourist gaze, many researchers have concerned themselves with both the production and consumption of the touristic images of brochures, media coverage and popular photography (Crang 1997). Semiotic analysis has been used by many tourism researchers for the analysis of the myths and fantasies that are created, sustained and refigured through both tourism marketing and tourist behaviour. The value of semiotic analysis is in the fact that it recognizes that there are usually several layers of meaning within any textual or visual analysis and that these are usually arbitrary but bound by particular cultural contexts. Thus, a semiotic approach encourages greater depth of analysis beyond the obvious or the literal to reveal the indirect and often unintentional levels of meaning in any text.
De Saussure (1916:16) originally defined semiotic analysis as 'a science that studies signs within society.' He argued that a sign is the result of the structural relationship between a signifier (a word) and the signified (object). As Echtner (1999: 48) illustrates, 'a sign might consist of a physical object, such as a sandy shoreline (signified), plus the associated signifier, the word 'beach' ... meaning is generated and communicated through the association between the signifier and signified ... ' Peirce (1934), meanwhile, further suggested that meaning is derived through a triadic relationship between the designatum (the object signified), the sign (the signifier used to represent the object) and the interpretant (the person who interprets the sign). Hence, a sign not only stands for something but also, crucially, stands for something to somebody (Hawkes 1977). As a consequence, researchers are asked to examine the relationships between the object and its representation, the representation and the interpreter, and the interpreter and the object (Fiske 1990 Peirce (1934) further placed signs into the three categories of icon, index and symbol. At the simplest level, an icon resembles the object signified in some way. For example, a photograph, postcard, painting or souvenir replica of the Taj Mahal ordinarily resembles the Taj Mahal itself and is thus an icon of the Taj Mahal. Despite the direct, apparently direct connection between sign and signified, however, all icons are grounded within particular cultural and historical contexts and may not mean anything to those that have never seen the object itself. At a deeper level of specificity, a symbol for Peirce has only an arbitrary relationship to the object in view. Words are a classic example of symbols that attempt to represent objects which, more importantly, become infused with arbitrary meanings (Wittgenstein 1953). For example, the words Taj Mahal may become suffused with symbolic meanings of love or religion depending on the interpreters' cultural and historical background (Edensor 1998). Used frequently once arbitrary meanings become fixed such that they become associated symbolically with a particular place or tourism destination itself such as India or, more specifically in terms of this example, the city of Agra. In contrast, the index has a direct causal relationship with its signified. For example, Echtner (1999: 49) argues that 'a suntan is an index of the amount of sun exposure. However, a suntan may also be used as an index of 'ruggedness' or 'healthiness". Again, understandings of the relationship represented by an index are culturally and historically proscribed and must be learnt. Crucially, it should be noted that icons, indices and symbols are not mutually exclusive. Thus images of the Taj Mahal may combine all three elements and also may mean many things to many people simultaneously. Barthes' (1984) work subsequently sought to uncover the ideological structures inherent within the semiotic analysis outlined above. He argued that there were various levels or layers of meaning, namely the denotative and the connotative. The denotative is the first level of meaning-the signs outlined above by Saussure. The second level is the connotative which Barthes considers as a type of myth. At the connotative level signs themselves can become signifiers thus creating whole chains of meanings or 'secret histories' . For example, Barthes (1984: 5) analyzed the Eiffel Tower as perhaps an example of a pure signifier to which a multitude of myths and meanings can be attached; Paris, love, enchantment, dreams and so on: 'it attracts meaning the way a lightning rod attracts thunderbolts'. Barthes' (1984) seminal analysis of The Blue Guide is also illustrative here: 'the travel guide functions as 'an agent of blindness' that focuses the traveller's attention on a limited range of landscape features, thereby 'overpowering' or even 'masking' the 'real' spectacle of human life and history and simultaneously providing an illusion of cultural stability and continuity' (Duncan and Duncan 1992: 20).
Although contemporary semiotic analysis does not follow a strictly defined procedure, it does, however, have a number of 'operational guidelines'. As Echtner (1999: 50) writes, the semiotic approach 'allows considerable analytical freedom and creativity in terms of research ... ' In terms of these operational guidelines a researcher should firstly, choose a representative and relatively closed corpus of data to analyze, secondly, segment the material to be analyzed into specific units and thirdly, develop an inventory of the key elements within the data set (as in content analysis). Following this the researcher should, fourthly, analyze the deeper meanings and compare these, and fifthly, provide a discussion of the cultural context in which the material is embedded. By following these procedures, the researcher's post-analysis will be in a position to produce a convincing and trustworthy argument.
Within tourism studies, a number of authors have analyzed tourism brochures from a semiotic perspective (Thurot and Thurot 1983;Uzzell 1984;Cohen 1989;Dann 1996aDann , 1996bMacCannell1989;Selwyn 1996;Cooper 1984). Thurot and Thurot's (1983) work argues that behind the obvious commercial realism the brochures justify an aristocratic lifestyle that the lower classes are encouraged to emulate for a short time period as tourists. Similarly, Uzzell's (1984) analysis of tourism brochures points out that photographs of wine represent the myth of a good life and symbolize a degree of freedom from everyday working life though the loss of inhibitions.
Whilst semiotic analysis has been primarily applied to tourism promotional materials, further research needs to be undertaken with longitudinal analyses of promotional materials particularly important in order to reveal the changing historical meanings of tourism promotion (see for example, Goss 1993) as well as the impact of internet based promotion (see, for example, Marcussen 1997). Semiotic analysis, though, has its limitations, most obviously because of its emphasis upon analyzing structures rather agency within societies. Semiotic analysts have also been criticized by post -structuralist researchers because of their failure to understand the unstable and power-laden nature of social meanings in a complex contemporary global society. Hence, we now move on to discuss the merits of post -structuralist discourse analysis for tourism research.

Discourse Analysis
Discourse itself is a nebulous term, attracting multiple meanings and understandings. However, from a post-26 structuralist theoretical perspective discourses can be conceptualized as the everyday social and cultural means through which people and agencies achieve action (Fairclough 1992). These actions include the construction, display and ascription of identities, social relations and know ledges, as well as the construction of the contexts in which actions can occur (Van Dijk 1993. The key difference that needs to be borne in mind when attempting discourse analysis is a rigorous awareness of the sociocultural context. Despite the multiplicity of meanings attaching to discourse, at least two distinct types of poststructuralist discourse analyses can be discerned. The first views discourses as being entwined with and as ultimately constituting larger formations of knowledge production. Discourse analyses from this perspective is concerned with critiquing the power of discourses. The second views discourses as socio-cultural texts rather like semiotic analysis. Discourse analysis here is concerned with the detailed deconstruction of both the meanings and the aporia within socio-cultural texts.

Discourse Analysis and the Production of Knowledge
From this perspective, discourses can be seen as ways of structuring areas of knowledge and social practice (Foucault 1972(Foucault , 1979. Hence, we might talk, for instance, about a discourse of medicine, a discourse of criminology or indeed a contested discourse of tourism research (see Tribe 1997Tribe , 2000. Following the work of Foucault (1965Foucault ( , 1977, discourses have been conceptualized as having a largely taken for granted power over social actions because of the ways in which they are able to acquire scientific status and thus credibility and legitimacy. Ultimately, specific discourses, as Foucault demonstrated, can lead to the creation of specific institutions. For example, discourses of psychiatry had the power to create certain types of institutions that we refer to as asylums, specific medical discourses led to the birth of the clinic, discourses ofleisure opened up the possibility of theme parks and so forth. Once such institutions, whether formal or informal (in the sense of conventions) are established, they further regulate human behaviour within the relevant contexts. Thus, in the same way that the operation of the institutions surrounding medicine continue to create and order medical subjects or clinical cases, tourist destinations, tour operators and travel writers regulate the world of the tourist to the extent of creating the tourist. The implications are that the nature of discursive knowledge production has an effect on what actions are undertaken and thus what outcomes are likely within any socio-cultural context. Within Foucauldian accounts, however, there is additionally a sense in which discursive contexts are constantly refigured and recreated through individual utterances, actions and texts. Thus,just as the tourist is created and controlled through engagement with the institutions of leisure and tourism, so those institutions are remade and further stabilized through those engagements. In exploring the regional heritage of northeastern Scotland, Knox (2001) uncovered the operation of power through the effects of a series of institutions that had come to shape that heritage in very particular ways. Folk song and literary traditions are regulated through the historical activities of ethnologists, and continue to be shaped by academics and enthusiasts through heritage societies, a specialist research centre and the activities of enthusiasts.
Discursive formations and their objects of knowledge have distinct genealogies that we can attempt to trace a route through to reveal the workings of power and the sometimes subtle shifting of meanings over time (Foucault 1972;Nietzsche 1998). Importantly, the connotation of specific words can vary significantly between moments in time, discourses and researchers. So, for example, the discourse of astronomy uses the term space in a different way from the discourse of geography. And, indeed as this paper has demonstrated the term discourse itself is similarly contested by various stakeholders. However, certain discourses will use a word in such a way that it becomes the dominant connotation whilst other discourses may use it in a subordinate or even subversive manner. Discourses constitute 'the limits within which ideas and practices are considered to be natural; that is, they set the bounds on what questions are considered relevant or even intelligible' (Barnes and Duncan 1992: 8). Such discursive formations can be thought of as broad frameworks for understanding and communicating that have gained some form oflegitimacy in society. For example, many tourism researchers have noted the development of specific tourism discourses and professional specialties such as management and hospitality (Cheong and Miller 2000;Hollinshead 1999;Coleman and Crang 2001;Tribe 1997Tribe , 2002. Here we might note the professionalization and institutionalization of tourism research through powerknowledge relationships (Hannam 2002). As Tribe (1997: 645) argues, 'Tourism has developed a network of communications which includes professional associations, conferences, books and journals'. He further points out, however, that 'tourism studies have not established anything that could be called a tradition that might impose its own unity' (Tribe 1997). Despite this lack of unity, when conceptualized as a discursive field the institutionalization of tourism research has allowed governments to often 'remove from the political realm problems which would otherwise be political' (Escobar 1984: 387 -388) and then to Tourism Recreation Research Vol. 30, No. 2, 2005 Discourse Analysis in Tourism Research: K. Hannam & D. Knox recast them into the apparently more neutral realm of tourism management for 'experts' to consider. These tourism experts arguably inhabit institutional environments as cosmopolitan intellectuals, members of a 'new tribe' of scribes who oversee the production and reproduction of tourism knowledge (see Watts 1993).
The institutionalization of tourism, meanwhile, takes place at various scales, ranging from the international tourism bodies such as the WTO, and national tourism planning bodies down to local level tourism development agencies and non-governmental organizations. These institutions are the agents of the deployment of tourism research, the network of new sites of authority which, taken as a whole, constitute the official apparatus of tourism development. This results in the dispersion oflocal centres of power-knowledge -the establishment of multiple sites of power and moral regulation that tourism researchers need to pay attention to.
As a result we might argue that various sociologists, anthropologists and geographers who have conducted tourism research have been successful to the extent that they have been able to penetrate, intervene, manage and control tourism practices on the one hand, as well as simultaneously professionalize, institutionalize and normalize strategies on the other, in increasingly detailed ways. From this angle tourism research could be perceived as not necessarily strategies that solve problems but rather as a field that is politically, economically, culturally and environmentally manageable for a variety of agencies and structures. Discourse analysis as in this context allows us to investigate the power /knowledge relations at the very heart of tourism research. Moreover, Foucault's (1977Foucault's ( , 1979 accounts of the regulation of the body and technologies of the self assuredly point us in the direction of more just simple textual analysis as they raise issues of the embodied social practices that are regulated by specific discourses. If individual texts need to be accounted for as part of wider systems of knowledge, it also needs to be acknowledged that the realm of discourse does not tell us the whole story of tourism, nor of any other cultural phenomenon.

Deconstruction
A second form of post -structuralist discourse analysis has also sometimes been referred to as a method of deconstruction. Some post -structuralist authors following Derrida (1977) have tended to argue that any discourse is the outcome of a series of, often incoherent, structures and utterances which are in conflict with each other. This form of post -structuralist discourse analysis seeks to develop a nuanced reading that unpacks in minute detail a particular text in the cultural context in which it is embedded. Drawing upon earlier semiotic work, deconstruction is not just interested in what is within the text itself but also in what has been left out and the 'secret' meanings that are not obvious . In contrast, Foucauldian discourse analysis has arguably tended to focus rather too much energy on the presences of the text to the exclusion of the absences or the silences (Billig 1997(Billig , 1999. Deconstruction, on the other hand, looks for the dichotomies that are written into any text 'and asks questions that tend to be ignored because of the dichotomy' (Feldman 1995: 52).
Collective rememberings or institutional knowledges require not only that memories and understandings are created in particular ways but also that other alternative know ledges are forgotten, repressed or denied. Whenever something is said, written or depicted there are, of course, always a number of alternative things that might have been said either as well or instead. Fundamentally, each and every discourse is intimately embedded within a larger system full of possible contradictions as well as possible supportive utterances. Derrida (1977) suggests that every discourse contains the conditions of its own undoing in that traces of other ideas or accounts have a tendency to present themselves to the reader or participant. This notion of the trace is concerned with the way in which the presence of one thing logically points to the absence or exclusion of its direct opposite (most obviously) or a wealth of other possibilities. It is also possible to think of these alternatives that fail to be actualized as utterances, practices or meanings as being constitutive of alternative discourses that fail to exert the same power over particular realms of practice. Deconstructive methodologies offer us the opportunity to explore alternative meanings and their potential effects alongside the ideas they contradict or fail to support.
For example, tourism researchers have used this form of discourse analysis in their critiques of heritage tourism.
Notions of heritage and inter alia heritage tourism are fundamentally contested from various social, cultural, economic and political points of view. As Graham et al., (2000: 2) note: ... heritage is a view from the present, either backward or forward to a future. In both cases, the viewpoint cannot be other than now, the perspective is blurred and indistinct and shaped by current concerns and predispositions, while the field of vision is restricted to a highly selective view of a small fraction of possible pasts or envisaged futures.
The historical narratives transmitted through heritage are selective, partial and frequently biased and distorted Qohnson 1999). The study of heritage tourism should thus involve acknowledging the contextual power relations that 28 underpin various representations and interpretations. Conceptualizing heritage as a series of contested meanings rather than as a series of artefacts allows us to analyze the conflicts and tensions (Graham et al. 2000). In this context Graham et al. (2000: 93)  Indeed, such heritage dissonance can lead to very real conflicts. Edensor (1997) outlines the reception of the film Braveheart in Scotland during the mid-1990s, pointing to conflicts surrounding not only authenticity but also the political uses to which a popular surge in nationalism might be put. Edensor (1997) illustrates how the symbolic spaces of the nation can become sites of conflict in his account of political rallies at Bannockburn during this period. Similarly, the landscape of Glencoe in the Western Highlands of Scotland is the site of a series of contested discourses relating to the Scottish nation, conflict and warfare, outdoor pursuits, beauty and serenity (Knox 2004). Alongside these competing discourses there is an alternative vision of that same landscape as imagined by local residents, keen to assert a legacy of everyday life and work. Through the operation of a folk museum, the production of numerous publications and public events, the people of Glencoe have sought to reshape their heritage at the expense of the story of the 1692 Massacre of Glencoe that tends to dominate heritage provision in, and accounts of, the landscape within which they live out their lives.
In some more extreme cases the memories associated with particular heritage landscapes can lead to those landscapes being actively erased or modified. This is most obviously illustrated by reference to the toppling of various Soviet era monuments in Eastern Europe following the fall of the Berlin Wall or the recent round of iconoclasm in Iraq witnessed in the aftermath of the second Gulf War. Even in these cases, however, there are always alternative interpretations and meanings. Today, in parts of Eastern Europe moves are afoot to memorialize, celebrate and reclaim the heritage of the communist era. These current developments can be allied to the growth of what some have pointed to as 'Dark Tourism' (Lennon and Foley 2000), the active remembering and consumption of death and disaster as a means of challenging overly-simplistic and romantic visions of the past. Lennon and Foley (2000) detail the development of heritage provision and interpretation at camps such as Aushwitz-Birkenau and Sachsenhausen to illustrate the processes through which a shift from the valourization of the nation and a celebration of anti-fascist values towards the commemoration of the holocaust and a monument warning of the dangers of anti-semitism have been effected.
Frequently, the techniques and philosophies of discourse analysis are utilized in an implicit fashion. What we are arguing for in this paper is a more sustained and critical engagement with both the materials under consideration and the methodologies through which they might be investigated. Discourse analysis as deconstruction allows us to examine the slippages and discontinuities of heritage dissonance in more depth. By paying attention to the ruptures and silences in any discourse of heritage, tourism researchers may reveal a more detailed and nuanced account of the complexities involved.

Conclusion
This paper has provided a pathway through a broad interdisciplinary range of research methodologies that are Discourse Analysis in Tourism Research: K. Hannam & D. Knox often conveniently grouped together under the term discourse analysis. The paper has differentiated between forms of content, textual, semiotic and discourse analysis and it has sought to demonstrate, inter alia, the usefulness of discourse analysis for the development of a more critical and reflective research perspective. Ultimately, forms of discourse analysis assume that all texts are produced intertextually in relation to other texts and that they are all embedded within power relations that give degrees of authority. Hence, rather than letting the words or images speak simply for themselves, forms of discourse analysis treat texts as mediated cultural products which are part of wider systems of knowledge which may set the limits for, or discipline, everyday life.
From this perspective, the discourses within which we operate largely determine what we enable ourselves to know (Shatter 1993). What we look for as tourism researchers have a more than minor preconditioning effect on the outcomes of our investigations (see Wittgenstein 1953). Thus tourism researchers should be interested not only in the discourses of tourism itself, but also in those that attach to the academic pursuit of tourism studies (Tribe 2002). In doing so, tourism research as an interdisciplinary pursuit has much to potentially gain in avoiding the narrative entrapments of particular academic disciplines through a critical and reflexive application of techniques of discourse analysis.