3.1 The Variable of the Context

Since the 1990s, Latin America has witnessed a surge in public policies, international development cooperation projects and locally driven initiatives to implement and replicate successful CBT models (^{Pérez Galán & Asensio, 2012}). Notably, the most significant characteristic of these proposals has been the low economic feasibility (e.g. ^{Cañada, 2015}; ^{Notzke, 2006}), but also the focal point of local conflicts (e.g. ^{Gascón, 2013}; ^{Schellhorn, 2010}). Beyond the management and design issues, one key aspect contributing to this low feasibility is the different outcomes due to local particularities. The rural contexts in which the CBT projects are developed are not always comparable. A paradigmatic example can be found in the neighbouring islands of Amantaní and Taquile, both located on Lake Titicaca. Both islands are inhabited by Quechua communities, with smallholder peasant populations. In the late 1970s, the residents of Amantaní began to promote the island as a tourist destination, adopting a model similar to Taquile Island, where tourism was successfully promoted by the local population a few years earlier (^{Zorn, 2004}).

The Taquile model viewed tourism as a community resource, with only the island-dwellers entitled to enjoy its benefits. Some residents opened small restaurants, while others profited as owners of motorboats, ferrying passengers from the city of Puno. However, the main source of revenue was accommodation, with nearly all island residents taking part in hosting tourists in their homes. A strict rotation system was implemented to distribute tourists among families. This revenue was added to the income obtained from the sale of textiles. By the early 1990s, the residents of Taquile had become a unique phenomenon in Peru: an indigenous middle class proud of its culture, demonstrated through their consistent wearing of traditional clothing, even when travelling to urban areas.

The inhabitants of Amantaní started their CBT projects adopting a similar model to the one used by Taquile: where tourism was considered a communal asset, and revenues were distributed equitably among members. In 1979, the Peruvian government officially recognised the island as an area of tourism interest. This endorsement instilled confidence among the islanders that a steady influx of tourists would ensue, emulating the success experienced in Taquile. Consequently, they undertook different actions to consolidate and leverage this new economic activity. Many residents fitted out a room in their homes to accommodate visitors, given the distance from the city of Puno, where the tourists came from, required overnight stays. The Town Hall was responsible for issuing accommodation licenses, establishing hygiene and comfort standards regulations, fixed pricing for full board and lodging (accommodation and meals), and implementing a rotation system to distribute tourists amongst the islanders fairly. Bolstered by governmental financial support, a craft hall was built to showcase and sell the islanders' handicrafts families (including textiles, stonework, basketry, etc.). However, within a few years, the hope placed on tourism as the solution to their financial problems faded away. Two elements explain this failure: the relatively small number of tourists and the monopolisation of tourist benefits by a minority.

Throughout the 1980s and 90s, the volume of visitors consistently fell below initial forecasts. While there was growth in visitor numbers between 1980 and 1990, they never reached noteworthy levels. At the end of the decade, an average of fifteen to twenty travellers arrived during the peak season of July and August, with limited tourism activity during other months. However, at the beginning of the 1990s, these figures dropped even lower due to various circumstances affecting Peru's tourism sector. These included the increase of the armed conflict between the Sendero Luminoso guerrilla movement and government forces, as well as the cholera epidemic in 1991. Moreover, the economic crisis from the previous decade posed challenges, particularly for the middle class as a target market for domestic tourism. However, by the mid-1990s, there were signs of improvement. The armed conflict became more localised to regions bordering the Amazon region, and the economy showed signs of recovery. Consequently, tourist numbers to Amantaní increased. By the mid-1990s, motorboat traffic increased significantly, with one or two boats transporting tourists daily during July and August.

Even with this increase, initial expectations were not met. This can be attributed to two main aspects. Firstly, demographic differences played a key role. In the mid-1990s, Taquile had a population of approximately 1,300 inhabitants, whereas Amantaní had nearly 4,000 residents. Amantaní required three times the number of visitors compared to Taquile to achieve an equivalent tourist-to-inhabitant ratio. Secondly, Taquile's proximity to Puno, the departure point for tourists visiting Lake Titicaca, absorbed most of the traffic, leaving Amantaní with fewer visitors.

Apart from the relative shortage of tourists, another element explains the disappointment in tourism among islanders: the unequal distribution of profits from tourism. Initially conceived as a communal resource exclusively for islanders, tourism was meant to provide benefits for the entire community. This understanding was codified in oral and written regulations, presenting communal ownership and management of tourism-related assets. However, empirical evidence has shown that community members are a long way from materialising an idealised, socially homogeneous cooperative society. Instead, community members are functional to the individual interests over collective welfare (^{Beltrán & Vaccaro, 2017}; ^{Gonzales de Olarte, 1994}).

In this context, the individual interest in profit distribution within the tourism sector was prominently shown in the case of Amantaní. Up to the mid-2000s, the island experienced a scenario where a minority, namely the motorboat owners, dominated the relatively low influx of tourists. These people lodged the tourists they ferried from the mainland in their own houses or those belonging to their close relatives. The netting principle was fulfilled, wherein individuals who owned specific assets (such as motorboats) were the primary beneficiaries of the common property (tourism). This concentration of tourism profits among the motorboat owners was facilitated, firstly, due to their ownership of the "means of production" concerning tourism, and secondly, by the control held by the motorboat owners over the governance structures, notably the Governorship. We will delve deeper into this topic in subsequent discussions.

The monopoly of tourism by the motorboat owners generated responses and conflicts. Opposition to the motorboat owners adopted different forms, with criticism being prominent. Graffiti often denounced motorboat owners and their leaders as thieves and their main leaders as thieves. Additionally, there were attempts to boycott activities aimed at promoting tourism, such as advocating for the return to traditional costumes as everyday wear. However, the most impactful action taken was the attempt to implement the initially planned distribution system for the tourists. This was attempted twice with the establishment of the job of a rotation system supervisor for a short period. The first instance was between 1982 and 1983. A few months after the implementation, the motorboat owners boycotted the distribution, alleging favouritism towards the supervisor's relatives. The second attempt was in 1993 when tourism in Peru started to rise again. The elected Mayor of the district of Amantaní spearheaded this effort, supported by the main opponents of the motorboat owners. In response, the motorboat owners reacted by accusing the supervisor of nepotism and the Mayor of embezzlement, leading to another boycott of the distribution system. Once again, the implementation of the distribution system failed. Overall, the attempts to challenge the tourism monopoly faced significant hurdles, primarily due to allegations of favouritism and corruption levelled against those tasked with implementing alternative systems.

The opposition to the tourism monopoly by the motorboat owners was occasional. Most of the islanders did not perceive any immediate benefit in persistently demanding a fairer distribution of tourists. The number was insufficient to cover the costs associated with equitable distribution among everyone, such as improving housing to meet minimum quality standards and keeping accommodation for tourists. Additionally, there was a concern that if motorboat owners ceased their transport activity primarily driven by tourism benefits, the island would suffer from worsened communication with the mainland.

The cases of Amantaní and Taquile offer thought-provoking approaches despite being similar communities. The two islands are rural areas that share the same ecosystem, a Quechua population, and an agricultural economy. Both communities are made up of a smallholding population, experiencing pressure due to land division, influenced by similar national and regional contexts, and culturally, socially and historically affected by their insular nature, with comparable tourism market. However, applying the same CBT management model, which considered tourism a common resource, led to opposing outcomes. By the mid-1990s, Taquile had become a cohesive society without any important socioeconomic differences, where the entire population took part in the new tourism initiatives and efforts to strengthen the community. In contrast, Amantaní experienced social and political conflicts arising from tourism, with benefits disproportionately benefiting a minority comprising less than 9% of the total island population.

3.2 The Paradigm Variable

A second factor contributing to the Dilemma of Duality refers to the theoretical framework employed. Generally, researchers approach the fieldwork with preconceived epistemological convictions that may predetermine the phenomenon's analysis. This does not necessarily imply distortion of reality but rather prioritisation of certain aspects over others. For instance, some authors argue that the Dilemma of Duality can arise when the plurality of agents involved in tourism activity is not considered. It is acknowledged that tourism side-effects do not affect all stakeholders in the same way (^{Blackstock, 2005}; Milano & ^{Gascón, 2017}; ^{Stronza, 2001}). Despite the direct association of tourism analyses with concepts such as culture, power, and identity, important perspectives such as the Foucaultian power-knowledge discourse and gender perspectives have often been overlooked (^{Salazar, 2006}). Likewise, much of the tourism research has been focused more on the host-guest paradigm rather than exploring the unequal power relationship between hosts (^{Sherlock, 2001}).

To illustrate the paradigm variable, the Pro-Poor Tourism (PPT) methodology helps as a relevant example. PPT arose as a strategy aimed at leveraging tourism as a tool to fight poverty. Under this framework, any tourism activity that contributes to the net income of the host community has been considered suitable, even if these earnings are marginal and disproportionately distributed among certain sectors of the local community (^{Ashley & Roe, 2001}). The PPT approach affirms that, despite appearing insignificant, the profits reaching the poorest social sectors hold significance within their constrained economies (^{Ashley, 2003}). This approach reflects a liberal perspective where poverty is viewed as an economic issue addressed through income growth (^{Torres & Momsen, 2004}). In this context, PPT would justify the development of tourism on Amantaní island in the mid-1990s, where, despite the concentration of profits among a minority, overall income for the island increased. Other population segments may have received marginal benefits, either by one-off sales of craftwork in the craft hall or by the trickle-down effect. Alternatively, in the worst-case scenario, segments of the population not benefiting from tourism would remain similar to pre-tourism development.

A relative understanding of poverty, such as the one proposed by Townsend (1970) and Sen (1981), markedly contrasts with the perspectives offered by PPT approaches. According to these authors, poverty stems from inequalities between social classes rather than merely from a lack of resources. This conceptual framework shifts the analysis of tourism's impact towards examining social stratification-whether tourism exacerbates or mitigates social inequalities-instead of solely focusing on the direct financial benefits tourism might bring. Consequently, within this paradigm, researchers will problematise whether the concentration of tourism revenues among a minority may affect the poverty levels of the rest of the population (^{Gascón, 2015}; ^{Harrison, 2008}).

Reflecting on the case of Amantaní in the mid-1990s, it is evident that tourism had increased social and economic disparities. To illustrate, during the peak months of July and August (high season), daily tourist arrivals ranged from 20 to 35, whereas the off-peak months saw a reduction to between three and six tourists per day. The cost for a tourist to travel to Amantaní was 6 US$, with accommodation and meals priced at 8 US$. Additional expenditures were incurred at local grocery stores and the Craft Shop, leading to an estimated annual revenue of approximately 54,990 US$ for the island.

A fair distribution of accommodation income across the local population would have meant less than 14 US$ per capita annually. However, the distribution favoured motorboat owners significantly. In 1995, fifteen boats operated in Amantaní, with a collective income of 1,500 US$ per motorboat. Given the total number of 90 motorboat owners, this resulted in an average income of 250 US$ per owner per year. While the income distribution among motorboat operators appeared fair - owing to an established rotation system that ensured a fair allocation of journeys throughout the year in high and low seasons - the reality of ownership and profit-sharing was more complex. The number of boat operators varied between one and eleven, and not all owners enjoyed equal rights or benefits. Specifically, some boat operators obtained higher profits from trip fares due to their larger initial investments in the motorboat consortium.

Therefore, most of the profits from tourism ended up accrued by motorboat owners, representing a significant concentration of economic benefits within a narrow segment of the community. From the perspective of PPT and the trickle-down effect, an analysis of tourism activity in Amantaní would appear positive despite the highly unequal distribution of profits. In the worst scenarios, a segment of the population might not receive any income from tourism. Yet, it could be argued that they did not incur losses in these instances. However, adopting a different perspective on poverty reveals that the societal unrest triggered by tourism could be considered a loss for the community's quality of life and health. Additionally, increased economic differences could undermine community governance mechanisms and have adverse political effects.

From the beginning of the tourism development until the late 1990s, the position of governor was predominantly held by motorboat owners. During this period, the Governorship was the most prestigious political role, characterised by significant ceremonial expenses and lacking remuneration. Therefore, only islanders with sufficient financial resources, predominantly those who operated motorboats, could afford to assume this position.

There were two reasons behind the motorboat owners' interest in controlling the Governorship. Firstly, it served to offset the opposition by some sectors of the community to their monopoly over tourism. Through this institution, they impeded any attempt to regulate the fair distribution of tourism accommodation. The motorboat operators' resistance to implementing a strict rotation system with oversight was due both to their control over tourist transportation and the political support afforded by the Governorship. Secondly, they sought to exploit the island's workforce (via community labour) and manipulate the Governorship to their advantage. Examination of the Governorship's minute book from 1975 to 1995 shows that most of the projects planned by this institution aligned with the interests of the motorboat owners, whether as motorboat operators (construction and maintenance of docks) or as the main beneficiaries of tourism revenues (campaigns for tourism promotion, establishing a craftwork festival to promote tourism, etc.).

In summary, while tourism increased the island's net income and visibly enhanced the economy and standard of living of a certain segment of the population, it failed to distribute benefits across the broader community. According to the PPT methodology, the tourism development of Amantaní might be deemed successful. However, alternative theoretical perspectives on poverty would likely offer a more critical evaluation. Such perspectives highlight that the increase in socioeconomic differentiation resulted in a power shift favouring the economically advantaged sector, thereby diminishing the influence of the non-beneficiary sector in decision-making processes.

3.3 The Methodological Variable

Richard ^{Butler (1980}) introduced the 'Tourism Area Life Cycle (TALC) concept, ' proposing that tourism practices undergo a series of phases characterised by growth and decline. This research highlights that such cycles extend beyond the temporal scope of the fieldwork, suggesting a need to consider temporal dynamics in understanding the outcomes of case studies in CBT projects. For this reason, the Dilemma of Duality and the disparity observed in CBT case study outcomes can also result from the timing of research efforts. Employing a longitudinal approach has been essential to capture the evolution of the Amantaní CBT initiative, englobing the various stages of the tourism activity lifecycle.

The tourism boom in Peru began in the mid-1990s, coinciding with the armed conflict in peripheral regions of the Amazon. This period saw an increase in the number of visitors to Amantaní, facilitated by the introduction of faster motorboats and the establishment of a new route via the Capachica peninsula. The increase in tourism during the mid-2000s triggered certain changes in the island's socioeconomic landscape. A key factor in this transformation was political: the Municipality of Amantaní emerged as a main political institution due to the increased funding by the Central State (^{Remy, 2005}). The Governorship, controlled by the motorboat owners, lost importance. Finally, the neoliberal policies implemented during Fujimori's government (1990-2000) ended the monopoly on lake transport to the islands, allowing foreign-operated motorboats to service the tourist route to Amantaní. The motorboats started to lose privileges during the term of the Mayor's office from 2011 to 2015 when some travel agencies from Puno shared out Amantaní's tourists using a rotational system.

Approximately 55% of Amantaní's population currently benefits from the rotational system. However, the distribution of profits remains very unequal, with people with a better infrastructure having exclusive agreements with major travel agencies. Although motorboat owners continue to transport tourists on the Puno-Amantaní route, their share of the tourism market is minimal, representing a minimal percentage of the overall visitor traffic.

Despite the positive impacts of tourism, a significant portion of the population remains excluded from hosting tourists. This exclusion stems from a variety of reasons. Elderly people often lack accommodations that meet the Ministry of Tourism's quality requirements necessary to qualify for subsidies from the Central State. Similarly, families benefiting from financial support from relatives abroad or those with income from agricultural activities may opt out of hosting tourists. Newly formed couples who are building their made-to-measure home tend to avoid participating in tourism. Additionally, some families spend most of the year away from the island. Finally, some islanders do not take part in tourism because they have other well-paid sources of income, such as grocery store owners or skilled tradespeople (master builders, bakers).

The enhancement in the life quality for the inhabitants of Amantaní is evident. An illustrative example is the two days per week fair organised on the island's main dock. The fair would have been unlikely in the 1980s and 1990s when the islanders' purchasing power was considerably lower. However, the increase in the quality of life cannot exclusively be attributed to the tourism sector. The economic upturn in Peru since the 2000s, following decades of crisis, has also benefitted the working-class and rural Andean communities, as indicated by macroeconomic statistics from the 1990s (^{Asensio, 2017}). New governmental aid programmes brought a small income to the low-income islanders, turning them into consumers. Moreover, the increases mentioned above in the municipal budgets have further contributed to the islanders' economic well-being, with many funds allocated to infrastructure projects that offered employment opportunities through paid labour within the traditional community work system. Finally, tourism growth also played a crucial role in a better island's economic standing.

The tourism boom has also had positive consequences on the farming ecosystem. In the mid-1990s, smallholding and dependence on foreign labour had caused the loss of agricultural borders and the soil quality deterioration. During this period, islanders often left to seek work between the harvest and sowing seasons (^{Gascón & Milano, 2018}), a time traditionally reserved for maintaining farming infrastructures. Amantaní, characterised by its steep terrain, employs a terraced farming system that dates back to pre-Hispanic times, requiring ongoing upkeep. Consequently, a significant portion of this infrastructure fell into disrepair. An NGO active in the early 1990s in terrace recovery estimated that around 25% of the growing land was eroded or lost (^{CIRTACC, 1991}). In the 2000s, the increase in family income due to tourism reduced the migratory trends and encouraged many former residents to return. Engagement in tourism implies two key considerations concerning the family workforce. On the one hand, tourism accommodation is not a consistent daily activity, with significant idle periods between guest stays, especially during the off-season. On the other hand, the nature of tourism work, requiring presence on the island to host visitors, makes it challenging to pursue external employment opportunities.

Farming income in the peasant economy has been reduced by the demographic increase in recent generations and even more due to the migratory trends over the past two decades. Despite these challenges, an initiative has been to expand agricultural boundaries, even onto low-yield marginal lands. For instance, the high pampa (plains), traditionally used for grazing, began to be cultivated in the mid-2010s. Although climate change has somewhat mitigated the impact of frost in the region, the land remains dry, rocky, and prone to strong winds and mini-tornados that can blow away the fertile topsoil. Finally, it is worth mentioning that most of the terraces have been recovered without any type of external assistance.

The development trajectory of Taquile Island has experienced the opposite process of Amantaní. By the mid-1990s, the model that had previously facilitated Taquile's integration into the tourism sector began to show signs of failure. According to ^{Zorn and Farthing (2007}), the main cause was the tour operators' capacity from Puno to disrupt the island's community organisation system. Entrepreneurs among the Taquileans established bilateral agreements with these tour operators, which were mutually beneficial arrangements: tour operators were able to lower accommodation prices and establish the conditions. In contrast, the islanders involved in these agreements saw an increase in the number of overnight stays. By the end of the 1990s, these practices had led to significant intra-community unrest and economic disparities, which only intensified in the following decade. In short, despite three decades of successful and cohesive engagement with the tourism sector, these internal challenges, exacerbated by an unfavourable external context, were insufficient to maintain the social capital the community had built up (Zorn, 2004). This case underscores the precarious balance communities navigate when integrating tourism into their socioeconomic fabric.

The literature on community-based tourism (CBT) often concludes from temporally limited research. Yet, as the experience of Taquile illustrates, tourism is inherently a dynamic and historical process that demands longitudinal analysis. Consequently, the assessment of CBT as either a success or a failure may vary significantly based on the specific temporal context in which the research is conducted. Understanding the Tourism Area Life Cycle within this longitudinal perspective is crucial for accurately evaluating the outcomes of tourism development in local communities

4.1 Discussion

Currently, the Variable of Context is recognised as a predominant factor in elucidating the Dilemma of Duality, gaining widespread acceptance. As mentioned above, comparing the cases of Taquile and Amantaní proves challenging due to their geographical, social, economic and cultural similarities. However, seemingly minor factors -such as population size or slight differences in proximity to the capital city- can lead to divergent outcomes from identical CBT tourism projects. The Variable of Context is inestimable as it avoids debates seemingly without any solution. By attributing duality to local characteristics and sidestepping potential ideological biases, it bypasses the dichotomy between optimistic and pessimistic views on tourism's role in rural areas.

While the Variable of Context is compelling and difficult to challenge, it does not negate the relevance of the other two variables. We argue that the theoretical and ideological framework employed by researchers (the "Paradigm Variable") plays a significant role in explaining the Dilemma of Duality. Scientific inquiry, particularly within the realm of social sciences and studies of social change, is inherently political (^{Habermas, 1970}). The once-prevailing belief in the axiological neutrality of science-especially within positivist traditions-has been largely contested in contemporary discourse. In general, tourism studies often move between two conflicting epistemologies. The first posits an infinite world with endless natural resources, suggesting that continuous economic growth can achieve development and prosperity. From this perspective, the goal is to elevate peasants' tourism-related income until it reaches a minimum threshold of well-being. The second acknowledges the finite Earth's resources, emphasising the importance of analysing the redistribution of profits and costs over mere net income increases (^{Gascón & Cañada, 2017}). Researchers ideologically aligned with the first paradigm may prioritise the gross income generated by tourism, overlooking the equity of its distribution. Conversely, those influenced by the second paradigm are more inclined to question whether tourism exacerbates socioeconomic disparities or if all community members have equal decision-making capacity in its management.

Finally, the Methodological Variable must be considered. Despite certain exceptions (e.g. ^{Escalera-Reyes & Díaz-Aguilar, 2017}; ^{Lee & Jan, 2019}), widely recognised theories such as the Tourism Area Life Cycle have not been extensively applied in CBT analysis. Neglecting the Tourism Area Life Cycle poses a problem when the ethnographic study merely captures a snapshot of a process that typically spans decades or even generations. However, it is common within academic literature to adopt a narrative style that portrays our analysis as a conclusive stage of the cycle. The oversight that this snapshot merely represents a phase within a long-term process presents a problem as it leads to biased conclusions.

4.2 Conclusion

Despite calls to leave behind a binary perspective in tourism analyses that classify the phenomenon solely as either a problem or solution (^{Cohen & Cohen, 2012}; ^{Xue & Kerstetter, 2019}), specialised literature continues showing a duality within findings. This dualism has been quite evident in the literature on CBT based on ethnographic research.

Ethnography, while a common method in anthropological research, has also found utility in other disciplines exploring CBT, such as geography (e.g. ^{Iorio & Corsale, 2014}), history (e.g. ^{Cañada, 2020}) and economics (e.g. ^{López-Guzmán, Sánchez-Cañizares & Pavón, 2011}). Ethnographic research usually delves into a circumscribed territorial context (^{Brewer, 2000}), facilitating a comprehensive analysis of the phenomenon under study. However, this methodology is not without its risks. One such risk is the tendency to generalise the functioning of CBT as a universal process, overlooking the nuanced elements and features of individual case studies. Moreover, ethnography's focus on a specific and limited timeframe may result in findings that present a snapshot of reality, potentially misleadingly perceived as static. In contrast, CBT is inherently a long-term process, characterised by phases that entail varying social, economic, and environmental impacts. Finally, the researcher's paradigm also plays a significant role in navigating the Dilemma of Duality. For instance, the case of Amantaní allows for the observation of diverse analyses and interpretations of rural poverty.

CBT research faces a challenging dialectic, which is tough to undertake. The research context, the researcher's theoretical framework perspective, and the limited duration of fieldwork are all factors that must be acknowledged. Nevertheless, some strategies can be implemented to mitigate these challenges. Firstly, it is necessary to recognise the presence of three variables previously identified in CBT analyses. One or a combination of these variables may account for disparities in research findings. Secondly, longitudinal ethnography and long-term research can provide a more nuanced understanding of CBT dynamics. Lastly, researchers must transparently articulate the theoretical framework and paradigm from which their case study has been analysed.