Through the agrarian lens: an extended approach to reflexive photography with farmers

This paper presents the documentation of the application of a reflexive photography approach and aims to elaborate its evaluation process. Farmers in Central Switzerland were first asked to take photos, after which interviews and group discussions were conducted with the farmers to explore their landscape perception. The reflexive photography method was applied in two steps. First, as the basis for the in-depth interviews evaluated with qualitative content analysis according to Philip Mayring, and second as the basis for group discussions evaluated with the documentary method according to Ralf Bohnsack. The paper’s goal is to provide deeper insight into the process of conducting a reflexive photography approach in rural studies.


VISUAL SOCIAL RESEARCH -AN INTRODUCTION
Within the realm of constructivism, qualitative research aims to extract different concepts of reality (Steinke 2007) by applying mostly inductive approaches to investigate social phenomena. Visual research methods are defined as the use of any visual material as part of the research process to investigate a research question. However, a range of methods exists; their diversity is inherent in both sorts of 'visual materials they work with and in the procedures to which those materials are subjected' (Rose 2014, 25). Originally developed by a sociologist (Harper 2012), the combination of photos and interviews constituting reflexive photography was adapted by Peter Dirksmeier (2012Dirksmeier ( , 2009) as a tool for spatial, qualitative social research in the field of geography. Similarities between reflexive photography and other visual qualitative research methods also demonstrate various combined approaches of photography and interview. In photo elicitation (Beilin 2005;Clark-Ibánez 2004) photos made by the researcher serve as the visual, whereas participants take the pictures themselves in auto-photography (Dodman 2003), photo-novella (Burke and Evans 2011), autodriven photo elicitation (Ford et al. 2017;Pauwels 2015;Schänzel and Smith 2015) and reflexive photography (Dirksmeier 2012). Whereas pictures themselves are analysed in auto-photography, the reflexive photography and auto-driven photo elicitation approaches aim to obtain more in-depth information in a personal interview. Nevertheless, the two approaches differ: photos are taken especially for the research process in reflexive photography, while already taken photos (for example, during holidays) serve as a stimulus in auto-driven photo elicitation.
Many publications based on visual studies focus primarily on the data survey and the process of picture taking itself, but often neglect the process of data evaluation (see Dirksmeier 2012;Beilin 2005;Robertson et al. 2016). For this reason, this study goes further with a pervasive evaluation process of gathering data with reflexive photography and conducting in-depth interviews. It then elaborates the idea of the approach and adds a next step: group discussion based on the same photos taken by the participants. While the individual habitus emerges from the personal interviews, the group discussions extract the collective habitus. Therefore, I briefly present the empirical and theoretical context to focus on the methodological application afterwards.

FARMING CONTEXT
Agriculture is still a main factor in shaping the appearance of the cultural landscape, even if the farming system has changed significantly with modernisation and the use of machines and automated work processes. The concept of multifunctional agriculture affords non-market benefits for society from farming (OECD 2001), which the Swiss population advocated in a referendum in 1996. As a result, the Swiss Constitution now includes landscape conservation as a new agricultural responsibility, along with maintaining decentralised settlements and natural life resources. With the reformation of Swiss agricultural policy, a system of agri-environmental schemes was introduced to meet the new multifunctional goals of farming (Norer 2005). Nevertheless, farmers tend to perform productivist farming, focusing on food production rather than on multifunctional goals (Wilson 2001), as this work contributes to their professional satisfaction and reputation (Stotten, Rudmann, and Schader 2010). Farm work is visible in the landscape and as such 'display[s] the farmer's knowledge, values and work ethic, and thus farmers appreciate tidy landscapes' (Rogge, Nevens, and Gulinck 2007, 160). However, the farming society values food production differently than landscape conservation work (Burton 2012;Burton, Kuczera, and Schwarz 2008;Warren et al. 2016). The new responsibilities of farming are especially important for the mountain areas, as the role of food production is limited in those less favoured areas. However, for the implementation of the agricultural policy and to fulfil the new agricultural task of landscape conservation, the perception of landscape among farmers plays a crucial role.

THEORETICAL CONTEXT
Landscape perception differs among people according to their individual experiences and knowledge. Olaf Kühne (2013Kühne ( , 2009 describes in the social constructivist landscape theory how the landscape is constituted as a social product. Furthermore, also the concept of habitus by Pierre Bourdieu (1977) is significant for landscape socialisation, as it defines a system of attitudes that are based on experienced history and are passed down through several generations of families and other social groups. Within the farming community, the habitus constitutes the definition of 'good farming', even though a diversity of habitus contributes to a diversity of 'good farming' practices (Evans, Morris, and Winter 2002).

SPATIAL FRAME AND SAMPLING
For the investigation, three municipalities in Central Switzerland were selected as case study communities within the Alpine area. Aspects such as number of inhabitants, altitude, total municipal area and landscape elements tend to be equivalent in those places, whereas other aspects differ significantly. Therefore, the selection aims to point out possible differences between regions, on the one hand, based on tourism, and on the other hand, based on nature and landscape conservation to extract the impact of ongoing local discourses on the landscape perception. Firstly, the municipality of Escholzmatt, part of the UNESCO biosphere Entlebuch, was selected as a case where nature and landscape conservation play a crucial role. Another location chosen was the municipality of Engelberg, which is a tourism destination for winter as well as summer sports. Thirdly, the municipality of Wolfenschiessen, where neither tourism, nature nor landscape conservation plays a crucial role, was selected.
Other empirical studies on landscape perception and preferences point out that human landscape perception varies depending on age, education, gender as well as personal or professional interests (Gao, Barbieri, and Valdivia 2014;Lindemann-Matthies et al. 2010;Strumse 1996). Therefore, those indicators guided the sampling, as well as farm size, form of occupation and mode of production. As nearly all registered farm managers in the selected municipalities were male and as fundamental differences in landscape perception due to gender are estimated, the sampling concentrated on male farm managers. For the other indicators mentioned above, the sampling aimed to ensure the variety among the 10 farmers in each municipality (see Appendix 1).

REFLEXIVE PHOTOGRAPHY
The reflexive photography approach, recently elaborated by Peter Dirksmeier (2013Dirksmeier ( , 2007, was applied in this study. It aims to visualise spatial experience for scientific research. According to him, this approach analyses the habitus, whereas Pierre Bourdieu (1993Bourdieu ( , 1977 did not mention any tool to be used to analyse habitus. Peter Dirksmeier (2007Dirksmeier ( , 2013 assumes that the object of the photo is marked by the photographer's habitus. The classical role allocation within the research processresearcher and respondentis therefore reversed, as the respondents unconsciously determine the research issue when taking the photos. During the reflexive photography interview afterwards, the respondent explains his/her motivation for taking the particular photo and thus expresses his/ her own habitus (Dirksmeier 2015(Dirksmeier , 2007. Therefore, the visual in this approach is seen as part of the research process itself, rather than as an autonomous research output (Pauwels 2010;Dirksmeier 2013).
In detail, the relational structure between the photo itself, the photographer and the observer (researcher) is entangled. As shown in Figure 1, several layers of relations exist among and next to each othersometimes these even overlap. Within the approach of social constructivism, a photo is an abstraction of reality seen through the eyes of the observer. With the process of taking the photo, the photographer creates Through the agrarian lens 375 a picture. Thus, the photo itself presents the photographer's idea of a picture rather than reality. It is the photographer's lens on reality, which at the same time is influenced by sociocultural aspects. Consequently, a photo shows less reality, but instead a subjectively culturally formed perception of reality. Therefore, the observer or the researcher plays a significant role in the process of reconstructing the original meaning of the photo. This meaning can simply be passed verbally to another person by the photographer (Overdick 2010). The approach of reflexive photography relies on the process of interpretation by the photographer and thus avoids suggested interpretations by the researcher (Pauwels 2015).
In order to apply the approach of reflexive photography, the first contact is made by phone or mail to obtain the person's consent to participate in the study. Disposable cameras are distributed either personally or by post accompanied by an explanatory letter. The disposable cameras mostly are equipped with a flash and can take up to 27 images. Even if the quality provided by disposable cameras is inferior to that of today's digital cameras, for the purposes of visual research as a process they are absolutely satisfactory (see Garrod 2008). Along with the camera comes written instructions for the first task to initiate the taking of photos. At any rate, the task itself is formulated in quite a vague manner, so that the test subject must ultimately decide which picture to take.
To avoid mishaps participants are sometimes asked to take photos of each object twice (see Stedman et al. 2004).
Experience shows that study subjects appreciate taking pictures, as it makes their research participation entertaining and playful (see Rose 2012), which might influence their willingness to participate. The taken images serve as a stimulus in the subsequent in-depth interview, thus making it easier for study subjects to talk about abstract issues and helping to bridge missing words. Narrative-generating interviews based on images generally take longer than those without (Collier and Collier 2004;Collier 1957). In addition, images tend to verbalise hidden aspects that would not have been mentioned without photos: Free associations to the pictured events (personal, political, historical) about the emotional state (consent, rejection, joy, sorrow, agitation, sentimentality etc.) up to a specific description of processes (working and manufacturing processes, traditional activities etc.) as well as the precise identification of pictured persons (name, status, profession, role, personality etc.), spatial structures (street names, buildings, institutions, political, ethnic and mental borders etc.) and additional image elements (designation of tools, machines and other artefacts). (own translation) (Overdick 2010, 212f) Finally, as the focus of the interview draws on the interpretation of the images by the photographer himself/herself, the content of the interview is directly chosen by the study subject as he or she becomes the 'expert'. Thus, the researcher has to be aware that also unforeseen issues come up during the interview, especially as the interviewees are not familiar with the

376
R. Stotten research criteria. Also, the additionally needed time, administrative and financial resources have to been considered in the research process (Dell Clark 1999).
The following in-depth interview (Chirban 1997) is based mainly on the photos and thus is less structured. The researcher often 'just' has to post clarifying questions. Photos and asked questions tend to encourage narratives; call-backs focus on comprehension and deeper interpretation of a statement (Lamnek 2010;Pauwels 2015).
Ethics codes serve to protect study subjects and researchers from liability issues during the research process. In the realm of visual studies, other aspects prevail: first, the study participant's consent to participate, the attempt to grant anonymity wherever possible and the copyright that remains with the photographer (Papademas and The International Visual Sociology Association 2009). One main issue is anonymity, which is difficult to achieve in visual research methods. Thus, any visual research process demands awareness of local circumstances, or as Gilian Rose (2012) states, 'each research project must devise its own ethical practices based on the specifics of its situation' (340).

Doing Visual Research with Farmers in Rural Switzerland
Data collection for this study was performed in summer 2012 and spring 2013. Decisive for the timing was the research process itself, but summer was chosen for data collection as there is much 'to see' in the landscape then, even though summer is a highly demanding season for farmers. In every municipality, a key person in relation to agriculture helped obtain access to the farmers. The farmers were asked for their consent by phone while at the same time being informed that the project is locally advocated. Motivation among farmers to participate was generally high.
Taking photos already started a process of thinking and awareness-raising concerning the research topic among the participating farmers. Afterwards, the photographs served as a stimulus for the in-depth interview, which focused on experiences and reflections on particular topics shown in the pictures (see Collier and Collier 2004;Pauwels 2010). At the same time, the photos were evaluated directly by the participant, namely the photographer himself/herself (Dirksmeier 2013).
To realise the reflexive photography approach, the study's farmers were given disposable cameras. In total, 10 cameras were distributed in each municipality, accompanied by a task description, an adapted project description, a statistical questionnaire and a postage-paid, return envelope. As it was difficult to reach the farmers (the data set for all farmers included only landline phone numbers), the cameras were sent to them over time. The task description instructed farmers as follows: This project takes a look at the cultural landscape and aims to investigate how farmers perceive it. Please be so kind as to take photos on your farm of what belongs to cultural landscape • this can be, from your point of view, aesthetically appealing elements, but also aesthetically displeasing elements.
Please take in total pictures of 8-12 motifs, and take each photo twice to avoid mishaps. Please send the film back to me within one week.
To limit the number of photos (as the disposable cameras had a capacity of 27 exposures) as well as to avoid mishaps, farmers were asked to take two photos of each motif.
Few farmers sent the cameras back within the requested time. After approximately two weeks they were called by phone and reminded to do so. In total, the return flow counted 27 cameras. One farmer in Wolfenschiessen took with my consent digital photos with his smart phone. Another camera was lost by the postal service (the farmer refused to receive a new camera and decided to terminate his participation). Another farmer was untraceableby phone or by mailafter consenting to participate. A letter sent in September 2012 informed him that the data collection had been terminated.
The films in the disposable cameras were successively developed and photo negatives were digitalised by a local photoshop. DIN A4 colour print-outs were made for the interviews. This process normally took three working days and interviews were held promptly. Farmers also appreciated that, because they could take the weather forecast into consideration for the appointment. Interviews were held on each farmer's farm and it was up to the farmers to include his partner and family or not, because even if it was the farmer himself who took the photos, the 'camera' was often discussed with the whole family.
A total of 28 interviews lasting 50-80 min were conducted with farmers on their farms between July and September 2012. At the beginning of each interview, photos were spread out on the table and the researcher tagged them in the order in which they were discussed by the farmer.
Through the agrarian lens

Evaluation
The interviews were subsequently transcribed from spoken Swiss German to written German and analysed using qualitative content analysis according to Philipp Mayring (2014). This systematic, rule-guided and theoretically grounded, step-by-step approach to qualitative text analysis is based on the inductive development of categories close to the given text material and the application of a deductive verification of those categories regarding the research questions and theory (Mayring 2014;Steigleder 2008). For the development of the system of categories, the computerbased tool ATLAS.ti was used to organise the text material. In total, 23 categories organised in seven dimensions were developed (see Appendix 2). In this study, this method of analysis served to extract the individual habitus.

THE EXTENDED APPROACH: PICTURE-LED GROUP DISCUSSIONS
Reconstructive social research-based group discussions aim to reconstruct the subject's atheoretical knowledge, as Karl Mannheim (1970) called it, which is better known as tacit knowledge (as termed by Michael Polanyi (1966)). This incorporated knowledge gives incentives for our actions, the habitus of our everyday practice. To reveal this knowledge, the composition of the group is homogeneous (for example, the same social group; for this study the farming society). The role of the moderator and the guideline of the discussion are reserved in character as more importance is given to the course of the discussion itself (Kühn and Koschel 2010). Also, the dynamics of the group and its discourse are significant for the understanding of common posture towards a phenomenon based on similar lifestyles and socialisation processes of the participants (Przyborski and Riegler 2010). The purpose of the introductory question is to start a self-running discussion. Issues discussed in the groups are thus determined by the participants and less by the researcher or moderator (Bremer and Christel 2007).

Group Discussions with Farmers
In a next step of this study, the same farmers were invited by letter to participate in a group discussion in their municipality in March 2013. Five photos from each municipality were chosen to ensure the greatest possible variety. A total of 15 photosspread out on the tableserved as stimulus for the discussion that was initiated by the moderator as follows: You might remember the personal interviews I had with all of you. Today, once again, the point of interest is cultural landscape and the question how farmers perceive, see and understand it. Farmers from three municipalities participated in this project.
Here we have photos made by farmers from all three municipalities. Maybe you could tell me something about what comes to mind when you see these photos.
The applied documentary method was mainly advanced by Rolf Bohnsack (2013a). It is based on the ideas of Werner Mangold (1960), who argues that in group discussions the 'speakers confirm, complete, correct each other, their statements build on each other; one could sometimes think that only one person is speaking, so closely does one debate contribution resembles another.
[…] The group opinion [respectively collective habitus] is not the "sum" of individual opinions, but rather the product of collective interactions' (49, own translation).
In the documentary reconstruction, the method aims to extract the collective habitus. To reveal the production of social phenomena, passages that show metaphorical density (dense depictions) and interactive density (high commitment) are defined for further investigation. Here, participants in the group discussion refer to their common tacit knowledge without any need for interpretation. The shared experiences among a group might be independent from each other; nevertheless, they are similar and lead to similar orientations (attitudes).
With the formulating interpretation, the researcher looks at 'what' is said by summarising the given statements. Working from this step, the next one, the reflecting interpretation, looks at 'how' something is said, which includes the researcher's interpretation. The discourse is analysed according to the nature of the statements: is it 378 R. Stotten a description, an argument or a narration? The discourse further considers how the discussion participants refer to each other (Bohnsack, Pfaff, and Weller 2010;Evers 2009). Already during this step, the comparison, either case-internal or case-crossing, starts. The ongoing creation of different sense-genetic types among the groups is based on the previous comparisons. First homogeneous structures are found in the discourse. Later, sociostructural aspects of each single group are also considered to identify collective orientations in a sociogenetic typecast (Bohnsack 2013b).
For each group discussion, selected passages were analysed with the formulating and reflecting interpretation. Thematically similar parts were extracted and compared in the comparative analysis (see Appendix 3).

INDIVIDUAL HABITUS: QUALITATIVE CONTENT ANALYSIS
Whereas a total of eight dimensions structure the text material from all the interviews (see Appendix 2), only one dimension is exemplarily presented here in part. The dimension 'general description of cultural landscape' includes statements about the general constitution of cultural landscape.

Wolfenschiessen
Farmer No. 1 in Wolfenschiessen considers landscape to be a changing image as it is a seasonal production cycle. It is influenced mainly by farming activities and thus it is the farmer who shapes it. This attitude includes not only responsibility for the landscape, but also the pride at being its designer, as the statement below demonstrates: Here it is when it was still standing, the fodder; when it was about two months old. And here it is cut. And then it's gone, right. You can't just let it be; landscape also changes as a result of our farming activities. In its essence it does not change, but its image changes. And, if one aims to preserve it, it means work, if you want to keep it the way it is right now. It doesn't stay like this by itself. (F 1)(see Figure 2) The cultural landscape for Farmer No. 10 in Wolfenschiessen is determined by its diversity as the embodiment of a diverse production. However, the diversity of mountain farming such as several small-scale elements and different places for grazing in mountain areas is only distinguishable among farmers themselves: When you go down to the level farming areas, where you find cultivation of corn and grain, there the cultural landscape becomes more interesting for me; the different elements. (…) Yeah, it has woods, it has grain, it has corn; it has different kinds of fields; also, the colour of the cultural landscape is more diverse there.
Here it is just green, several kinds of green. We, the farmers, notice that this is a highfertility meadow and that is a low-fertility meadow; as a farmer, you can recognize that from the colour of the meadow. (F 10) ESCHOLZMATT For Farmer No. 14 in Escholzmatt landscape is made up of the existing small-scale structured farm land, including trees, hedgerows or tarns. Animals, as for example cattle, are not only seen as being part of the landscape, they are also 'landscape conservationists' (F 14), which means that livestock farming itself has a positive impact on landscape conservation. Farmer No. 18 mentions that landscape is not shaped by the local population everywhere; a distinction is made between different regions. Human influence on the presence of cultural landscape is explicitly highlighted and differs in regional characteristics of farm management applied by local farmers, as demonstrated in the following comment: Cultural landscape, that's a landscape shaped by a generation of people. And those people belong to the landscape und determine how they want the landscape to be. They built it up like that. But that doesn't work everywhere. (F 18) Farmer No. 11 relates to his ancestors in his construction of landscape. Therefore, elements already installed by their fathers or grandfathers contribute to a positive appreciation of the farm landscape as an act of mutual respecteven if those elements are more likely to represent an obstacle for farming.   Figure 4) The Agrarian Lens: results of the Interviews In the interviews, the farmers express their individual habitus that is characterised by the attitude that a farmed landscape is valuable. From this, farmers are able to acquire symbolic capital or reputation as their work becomes visible in the landscape (see Burton 2012). In the following, exemplary aspects constituting the individual habitus are presented for each municipality.

Individual Habitus: Wolfenschiessen
The individual habitus in Wolfenschiessen is shaped by farming management processes that change seasonally. Therefore, as the cultural landscape is shaped by farming, human beings, as well as animals, belong to its concept. A diverse landscape is thus the result of the farmers' own actions undertaken to meet his moral values.

Individual Habitus: Escholzmatt
In the municipality of Escholzmatt the individual habitus is strongly influenced by human activity. Landscape represents the cultural heritage of the local farming society and must thus be conserved. The farmers' construction of cultural landscape is based on its function as the habitat of the local population.

Individual Habitus: Engelberg
In Engelberg farmers construct their individual habitus of landscape as an opposite pole to natural landscape. Cultural landscape reflects the human livelihood. Thus, its perception includes human action as well as settlement area and tourism infrastructures.

COLLECTIVE HABITUS: DOCUMENTARY METHOD
The sense-genetic typecast, which is based on the formulating and reflecting interpretation, has been elaborated to further formulate the socio-genic types that also include the socio-economic context of the cases. The following short transcript passages from the sense-genetic typecast show an example of a formulation of the socio-genic type.

Wolfenschiessen
Passage I (see Figure 5) Subsequently, Farmer No. 6 adds that cultural landscape represents a certain freedom that demonstrates a positive field of vision. However, this field of vision is limited by the requirement that this free space must in future continue to contain the diverse elements that it now contains. Furthermore, Farmer No. 10 underlines that the photo shows the diversity of cultural landscape including machines. This statement is validated by Farmer No. 5. In the further elaboration, Farmer No. 10 also includes traditional elements in what he constitutes as cultural landscape, such as an old hay barn. This highly interactive part of the group discussion documents that, on the one hand, the state of the current cultural landscape is incorporated as the target state. On the other hand, farmers are aware that the cultural landscape continues to undergo change, and thus is dynamic.

Escholzmatt
Passage II (see Figure 6) with two buckets and I was able to at least put a couple of pears into them. It took three trees to just fill them. F 12: That was an exquisite crop. >ironically< F 11: The other one I filled from three trees, apples.
Yeah. F 12: Yeah! Farmer No. 11 argues that, on the one hand, ‚one' wants things to be ecological, while, on the other hand, asks himself why he has the trees. Farmer No. 12 quickly validates this statement and Farmer No. 11 elaborates that the trees exist for their blossoms and for the birds. He does not finish the subsequent sentences, so that Farmer No. 12 argues further that trees do not provide a regular income. Farmer No. 11 states that he got not even two buckets of pears from three trees, which Farmer No. 12 validates as a modest return. This part illustrates that trees belong to the cultural landscape. However, their returns are not primary, but instead intangible in value.  F 28: With the silage bales, we farmers can make a major impact. Well, this one is close to the stable, more or less. You know, it's worse in places where it just lays around all winter. F 22: Or the snow-white ones. Those are even more visible in summer. F 25: That's to make them easier to count, you know. >ironically< F 22: Yeah, that's true.
Farmer No. 21 begins the discussion by wondering where the orchard is. Then, Farmer No. 22 states that there is no orchard as no fruit is cultivated in the area. Furthermore, Farmer No. 24 adds that it could be in 'Urnerland' (Canton Uri), which is validated by Farmer No. 25. Thereupon, Farmer No. 28 mentions that the depicted stone wall is located at another farmer's farm in Engelberg, which is validated by Farmer No. 22. Farmer No. 24 explains further that that is the new stone wall built by the farmer himself, and Farmer No. 28 claims that it also belongs to the cultural landscape, even if it is still new and has yet to become an element of cultural landscape. In this part, the opinion is

384
R. Stotten expressed that old traditional elements are considered to be cultural landscape, whereas this is not necessarily true of new elements, which must first become cultural landscape.
Farmer No. 22 further states that he took photos of the telephone pole and the silage bales to demonstrate negative aspects of cultural landscape. He further explains that those elements also raise debates.

Sense-and Socio-Genetic Typecast
Wolfenschiessen Farmers in Wolfenschiessen have a concept of landscape shaped by the present as well as by ongoing processes from their cultivation. Thus, landscape is not seen as a target state. Landscape further represents a link between their own lifestyle and space, as cultural landscape is an expression of moral values regarding a person's lifestyle, especially in relation to family and friends. Farmers recognise their own moral attitudes, which respect traditional values, in the landscape as the opposite of urban lifestyle values. With the farm work performed daily the awareness for landscape issues increases. In this way farmers learn how to read and understand the landscape. Landscape also represents a link to the past, and maintaining the landscape is also a sign of respect towards their ancestors.
Even if Wolfenschiessen is neither part of an official park nor does it benefit from massive tourism influence, farmers here show a cooperative attitude towards other landscape players.

Escholzmatt
Farmers in Escholzmatt demonstrate a concept of landscape that highlights an image shaped by the former local population. Modifications, even if caused by farming activities, are seen rather negatively.
Innovations are incorporated into the landscape concept slowly, which is therefore seen to be rather immobile and static. It is also strongly politically influenced, and a valuable landscape is equated with ecological compensation areas, as demanded by Swiss agricultural policy. Landscape is also visual evidence of good farming practices. Within the farming society, good farming skills are recognisable in the landscape and therefore it is a profiling platform.
The attitude towards other actors involved in the landscape, and especially towards the UNESCO Biosphere Entlebuch (UBE) as a leading public body in the region, is critical. The farmers present the UBE as a construct with only external impact, and without any impact within the region. Nevertheless, communication among landscape-related actors within the UBE is seen positively and causes agro-environmental goals to be incorporated into farming.

Engelberg
Farmers in Engelberg construct the term landscape predominantly with traditional farming elements. However, modern farming elements are excluded from this construction or are viewed negatively. Therefore, the landscape is a reflection of the quality of farm work performed by a farmer. His moral work values expressed as neatness and tidiness are visualised in the landscape.
Farmers are aware of the monetary value of the landscape for tourism in Engelberg, as many farmers FIGURE 8. Types of landscape perception. Source: Authors earn additional income from the tourism sector (e.g. by granting permission for tourists to cross their land on the slopes, but also from performing additional work in the tourism sector).

Types of Landscape Perception: results of the Group Discussions
Among the communities investigated, landscape perception differs on two levels: on the vertical axis the difference in perception differs depending on whether it is shaped by self-determination (individual aspects are important) or is externally determined (through agricultural policy, tourism …) (see Figure 8). On this level landscape, perception is either led intrinsically or extrinsically. Horizontally, landscape construction differentiates between a dynamic landscape as a process and a static landscape as a rigid conception. Different forms emerged among farmers participating in this study: in Wolfenschiessen farmers express a self-determined perception, as they are not influenced by the political or economic discourse in the municipality. By contrast, farmers in Engelberg and Escholzmatt demonstrate an externally determined construction of landscape. In Engelberg landscape is closely linked to tourism, whereas in Escholzmatt it relies on concepts of nature and landscape conservation. Also, on the level of dynamics, the communities show differences: farmers in Escholzmatt express a static construction of landscape based on previous ideals. By contrast, farmers in Wolfenschiessen and Engelberg manifest a dynamic understanding, including work processes as well as activities undertaken by other actors in the landscape.

REVIEW OF THE METHOD
Triangulation of two (or more) qualitative research methods aims to deepen or complement empirical learning (Flick 2010). With the combination of reflexive photography-related in-depth interviews and group discussions the investigated social reality gives comprehensive findings. Whereas the in-depth interviews encourage narratives about everyday practices, the group discussions in a homogeneous cluster show the production of social phenomena.
Anonymity is difficult to achieve in visual research methods, as discussed above. In this study where the photos printed in publications permitted local people to easily identify particular farmers. For this reason, farmers granted their consent for each photo to be published. Farmers participating in the study highly appreciated the data collection process: first for the purpose of initiating a process of awareness raising and reflection on such an abstract issue by taking photos on their own, and subsequently for the purpose of interpreting their own photos in a personal interview and finally discussing them with other farmers. This process enabled farmers to substantiate their personal attitudes towards landscapealso as a matter of personal or professional development.
Being on the go with the camera has made the farmers aware of their everyday landscape and intensified their view of the cultural landscape. One farmer expressed his reflexion about the process of making photos in one picture (see Figure 9). It visualised nicely that the task to look at the landscape through the agrarian lens was perceived like seeing it through a mirror, which enabled a conscious perception.
As mentioned by one farmer in Escholzmatt, it has to be taken into consideration whether photos taken might also have been influenced by an ongoing discourse on agricultural policy. Also, Thomas Overdick (2010) questions whether the taken photo is a representation of the environment and experiences made or whether it is instead a reactive and obsessive application of internalised rules. Nevertheless, interview material provided much material resulting from the habitus. Even if social desirability might be expressed in topics related to agricultural policy, I argue that most of the attitudes shown in the interviews and group discussion do not fall in the realm of any issue of social desirability. However, qualitative social research is not representative and quality criteria are not reliability, validity or representativeness, whereas quality is proved by plausibility and the elimination of arbitrariness (see Leibenath 2013). Therefore, this paper aims to contribute to the intersubjective comprehensibility of a visual qualitative research process.
The open character of the reflexive photography method enables participants to integrate for the researcher unexpected issues. For me, as a researcher it was surprising during this study that farmers took several photos of their families and friends. Therewith they intended to refer to the importance of human beings to perform the labour demanding work to maintain such a landscape (see Figure 10).

CONCLUSION
This research project considers the approach of reflexive photography as a tool for triggering in-depth information.
The photos shown during the interviews encourage visual descriptions of what social practice is with regard to cultural landscape. Reflexive photography not only enables the researcher to obtain research data, but also enables the farmers to critically reflect on their everyday practice in the landscape. It thus allows researchers to leave a positive benefit among the study subjects and goes further than a simple data collection process.
Whereas other visual research methods such as photo elicitation using photos taken by the researcher or autodriven photo elicitation relying on already existing photos taken by the participant, the extended method of reflexive photo presented here activates the participants capacity to unconsciously determine the research issue from the very beginning and through different steps. This facilitates the researcher to integrate verbally transmitted induvial meanings of the photos to extract the individual habitus based on everyday practises (personal interview and qualitative content analysis) and further to carve out the collective habitus represented in social phenomena (group discussions and documentary method). Therefore, I argue that an extended method of reflexive photography is suitable for explorative social constructivist research to enable the consideration of participants' determination of the issue under investigation.
Farmers participating in this study construct landscape out of diverse aspects in their everyday life that are primarily based on agricultural cultivation. The individual habitus is strongly expressed in moral values that were transmitted during the primary landscape socialisation. Those moral values are, for example, seen in work related to respecting ancestors who made the land arable. In contrast, the collective habitus of farmers in this study is expressed in their attitude towards ongoing public, local discourse. Within this collective habitus, the habitual practices in Engelberg are influenced by local tourism, whereas nature conservation is more at the focus of the farmers in Escholzmatt. The consideration of those values shaping the landscape perception among farmers is of importance for the implementation of incentives for landscape conservations schemes.