Transforming Burundian “taste of place”: From shunned in commercial blends to specialty coffee

ABSTRACT

Human geography has a history of engaging with place-based-quality products through a variety of concepts such as terroir, geographical indicators (GIs), and fictive places. While the efforts necessary to construct a “taste of place” have been explored, it remains unclear how a “taste of place” is established, and by whom. The article explores how relations between quality, products, and places are produced on the ground. More specifically, it addresses the question of what it takes to reconfigure a “taste of place.” Based on ethnographic fieldwork carried out in 2017 concerning a coffee producer in Burundi, the article shows how Burundian coffee was reconfigured from an inferior commodity coffee to a sought-after specialty coffee. The findings show that reconfiguring “a taste of place” requires both material and symbolic quality attributes. By underlining the importance of material quality attributes that are place-dependent, it provides a different angle to the discursive approach to “taste of place” in human geography. The author concludes that creating a “taste of place” requires taming space into a consumable representation of place through discursive and material practices.

KEYWORDS:

• Burundi
• fictive place
• specialty coffee
• “taste of place”

Introducing “taste of place”

The identities of some places and products, such as Champagne, Darjeeling, and parmesan, have become so intertwined that they are synonymous. The ways that the quality of agricultural products are shaped by the place of their production have been central to the study of wine (Demossier 2011; Hill 2022), cheese (Turbes et al. 2016), tequila (Bowen 2010), water (Jones 2009), beer (Bråtå 2017), and coffee (Smith 2018; Williams et al. 2022). However, just because a good is produced in a specific place does not classify it as having a positive quality. It is difficult to imagine products such as dried mushrooms or salt from areas associated with nuclear disaster being associated with quality.

The links between products, places, and quality have been addressed in the literature in two broad strands, one more practically focused and the other more theoretical. Some authors explore practical place-based concepts, such as terroir and geographical indications (GIs), and assert that links between places, products, and quality do exist (Barham 2003; Josling 2006; Pike 2009; Besky 2014). Discussions on whether GIs should be regulated dominate the literature (Barham 2003; Josling 2006; Defrancesco et al. 2012; Besky 2014; Hill 2022). A more theoretical engagement of links between products, places, and quality can be found in publications on geographies of consumption (e.g., Goodman et al. 2010; Mansvelt 2012). The central focus of this literature is how consumption constructs fictive places (Cook & Crang 1996; Overton & Murray 2016). The placeness of products is important in both strands of literature, but the actual process of configuring/reconfiguring the “taste of place” by producers is less prevalent in the literature.

Based on fieldwork in Burundi, this article examines how a “taste of place” was transformed there. Burundi seemed an unlikely place to be known for specialty coffee, the main reason being the high prevalence of a Potato Taste Defect (PTD) caused by local bacteria and pests (Hale et al. 2022). Burundian coffee is considered an “ugly duckling” among coffee origins, due to the coffee with the PTD tasting and smelling like raw potato in liquid form. Despite the risks of PTD, Burundi has ideal biophysical conditions for growing specialty coffee (Borrella et al. 2015; Lenaghan et al. 2018; Smith 2018). In this article, I document how a “taste of place” was reconfigured into signifying quality in a place known for negative quality attributes. The data were derived from ethnographic fieldwork in Burundi, where Mwiriwe¹ Coffee, which aimed to set Burundi on the specialty coffee map, was studied.

This research adds to what is known about the links between place, quality, and products, and specifically how a “taste of place” is configured/reconfigured. It raises questions about the view that a “taste of place” is a fictive place constructed by consumers, connoisseurs, and producers (Overton & Murray 2016). This article suggests that a relational approach allows for a better understanding of material and discursive aspects of place-making by producers (Massey 1994; 2005; Cresswell 2015).

The article is structured in the following way. First, a brief review of the literature addressing a “taste of place” is provided. Thereafter, Burundi is introduced as an origin of coffee, and the methods used to study coffee production in a few of the thousand Burundian coffee-producing hill sites. The presentation of the findings is divided into three sections, each of which addresses an important part of the global value chain: farm level, washing station, and buyers. Each section presents material and symbolic quality attributes found necessary to reconfigure Burundian coffee into a “taste of place.” This is followed by a discussion of how taming the space of coffee production into a “taste of place” was achieved, including a reflection on spatial inequalities of power. Finally, the Conclusions section is presented.

Geographies of “taste of place”

The cultural concept of terroir is central in quests to understand how place defines the quality of agricultural products (Trubek 2009; Demossier 2011). The word terroir is so imbued in French culture that it cannot be translated meaningfully (Trubek 2009). This mysterious connection between terroir and “taste of place” is best portrayed by Trubek (2009, 6–8):

In the act of tasting, when a bite of food or a sip of wine moves through the mouth and into the body, culture and nature become one […] If the taste is produced by place, how does it work? Can earth, air, sun, and water really make such a powerful imprint in my mouth?

Terroir has historically referred to specific climates, minerals in the soil, and physical geographies that result in specific tastes of agricultural products, such as wine (Gade 2004; Trubek 2009; Hill 2020; Overton & Murray 2021). For example, one study found that Cheddar cheeses made with milk from farms within a 5 km radius were grouped together, while cheeses made from milk originating from farther away were perceived as different (Turbes et al. 2016). This suggests that the flavor of a product is directly tied to a demarcated geographical location, a place with a concrete form (Cresswell 2015).

The human element, with local histories and specific agricultural and processing practices, was later acknowledged as central to terroir (Demossier 2011; Hill 2022; Williams et al. 2022). Today, terroir is no longer thought of as a timeless geographical location, but as a vibrant, constantly changing discursive spatial concept that links histories, practices, actors, and social organizations together (Demossier 2011; Overton & Murray 2021; Hill 2022; West 2022). Studies of New World wines that actively construct new places of quality have shown that the links between places, products, and quality are not intrinsic, but are discursively created (Murray & Overton 2011; Defrancesco et al. 2012; Overton & Murray 2014; Mathews & Brasher 2016; Parga-Dans & Alonso González 2017; Rytkönen et al. 2021).

Following on from the above discursive trajectory, a number of studies have engaged with terroir based on Lefebvre’s social construction of space (Overton 2010; Murray & Overton 2011; Weiss 2011; Lefebvre 2013; Overton & Murray 2016; Bråtå 2017). One study, based on ethnographic evidence from pork production in North Carolina, revealed how the “taste of place” was constructed by chefs and consumers at farmers’ markets (Weiss 2011). In a similar manner, studies of the process of fine wine qualification, realized in New Zealand wine markets, have shown how devices such as magazines, guides, and rankings have played a key role in co-constructing a global fine wine market (Rainer 2021). These studies have shown how the consuming end of the value chain, through the acts of discernment, were important discursive practices in constructing a “taste of place” (Weiss 2011; Overton & Murray 2021; Rainer 2021).

Terroir and a “taste of place” are deemed by some geographers to be social constructions to such an extent that they are suggested to be fictive places (Overton & Murray 2016). Fictive places are considered a factor of production in of itself, used for capitalist accumulation and exchange. A fictive place is defined as follows:

it builds on [a] material basis for land and social perceptions of place to construct a particular set of imputed values around a legally defined and bounded entity that acquires distinct economic value. This value is imaginary until it is made real by the purchase of a place-based commodity at a premium price by consumers: the ‘consumption of space’. (Overton & Murray 2016, 806)

The definition of fictive places includes both material and discursive construction of places (Overton 2010; Murray & Overton 2011; Overton & Murray 2014; 2021; Bidwell et al. 2018). However, the limited range of products and places associated with “quality” highlights that the place-product link is not infinitely malleable. Thus, constructions of fictive places have been confined by factors that are not fully understood.

Quality products with a “taste of place” are increasingly in demand globally (West 2022). This is especially the case for coffee, for which the market share of specialty coffee from a single origin has been growing since the 1980s (Roseberry 2005; Bro & Clay 2017). Studies of geographies of consumption have shown that people pay more when more specific geographical information is provided (Pike 2009; Teuber 2010; Overton & Murray 2021). For example, a study conducted in New Zealand found higher prices associated with the more detailed place of origin and suggested that consumers perceived higher quality as a result (Overton & Murray 2021). However, the focus has been on the role of consumers in construction of space and place: “The relationship among consumption, space and place can indeed be a mystification since attention is often paid to appearances, rather than materiality, and both historical and contemporary analysis fails to capture the social forces that actually produce space” (Goodman et al. 2010, xii).

The efforts of producers that are necessary to create fictive places have been addressed previously (Ponte & Gibbon 2005; Pike 2009; Rainer 2021). One study asserts how the collaborative efforts between humans and yeast was central in establishing a regional expression of wine and qualifying the product as fine wine (Rainer 2021). Both material efforts (Rainer 2021) and discursive efforts (Weiss 2011; Overton & Murray 2021) have been found crucial for the construction of a “taste of place,” but what is still unclear is how the relation between quality and a “taste of place” is established (Parga-Dans & Alonso González 2017; Rainer 2021). Since the actual process of configuring a “taste of place” on the ground is not fully understood, I focus on it in this article.

When working with a “taste of place,” “local is not simply an existential condition of being in a place, it is a specific orientation to how space is produced” (Weiss 2011, 456). Space and place need to be distinguished in order to analyze the making of a “taste of place” by producers. In this article space is understood as a set of ongoing and interconnected relations producing coffee that includes global demand for specialty coffee, climate change, the local microflora, and actors across the whole value chain (Massey 1994; 2005). In this article, place refers to the locality of the specific Burundian hills² where coffee is produced, and where the sense of place is given meaning through subjective experiences (Cresswell 2015). Places, being both objective and subjective, require an analytical framework that looks at both at the material and the discursive aspects of place-making. A relational approach to space as material-discursive practices intersects well with the material and symbolic attributes of coffee, and in this article it makes up the analytical framework for the “taste of place.”

What is special about specialty coffee?

It is useful to think of the making of specialty coffee as a process that honors coffee as a fruit. Rather than producing a standardized product, all processing is aimed at enhancing the unique quality of the fruit (L. Rosenberg et al. 2018). The Specialty Coffee Association (SCA) categorizes coffee as “specialty” when it scores 80–100 points (Table 1). Skilled cuppers use smell, taste, and sight to calculate the quality of the coffee (more points means higher quality). Cuppers are essential actors in qualifying coffee as specialty coffee and making the connection between the taste and place, similar to how sommeliers appraise quality in the wine industry.

Table 1. SCA quality categories (source: Specialty Coffee Association 2003)

Score	Grade	Specialty quality (Yes/No)
90–100	Outstanding	Yes
85–89.99	Excellent	Yes
80–84.99	Very Good	Yes
<80.0	Below Specialty Quality	No

Coffee has two types of qualities that are crucial to understand in the making of a “taste of place”—material and symbolic (Daviron & Ponte 2005). Material quality attributes are embedded in the product, which can be “objectively” measured using human senses (vision, taste, smell, hearing, touch) and/or technological devices (Daviron & Ponte 2005; L. Rosenberg et al. 2018). Symbolic quality attributes are trademarks, certifications, and geographical indicators, and they are detached from the material quality of the product. A “taste of place” is therefore a highly relevant symbolic quality allowing for the consumption of place (Daviron & Ponte 2005; L. Rosenberg et al. 2018).

An SCA cupping form is designed to evaluate the material quality attributes of aroma, flavor, aftertaste, acidity, body, uniformity, clean cup, balance, sweetness, and absence of defects. The SCA quality score is intended as a guarantee of the material quality attributes, as only coffee high in material quality scores above 80 points (L. Rosenberg et al. 2018). Symbolic qualities can be achieved without material quality attributes, such as fair trade, single-origin, or organically certified coffee (Daviron & Ponte 2005; L. Rosenberg et al. 2018). However, specialty coffee is contingent on outstanding material quality attributes that are assessed by cuppers.

What makes a material quality? The material quality attributes of coffee are place-bound and generated by inherent factors, such as (1) the genetic type of the coffee tree (e.g., Coffea arabica, Coffea robusta), (2) the cultivar (e.g., Bourbon, Typica, Geisha), (3) agro-climatic conditions (place-specific soil type, altitude, micro-climate), and (4) external factors such as farm practices, bean harvesting procedures, primary processing method (washed or natural), export processing (milling), and handling and storage during the journey from origin to place of consumption (Daviron & Ponte 2005; Specialty Coffee Association 2022). These attributes of material quality are significant because they guide the material and discursive practices of making a “taste of place.” Furthermore, the material quality needs to have minimal defects (maximum of five defects per 300 g of coffee), because specialty coffee should have clean distinctive flavors that can be traced to a specific regional terroir (Smith 2018). This means that the place-based symbolic quality of the coffee is directly entangled with the material quality in the specialty market segment.

Compared to most other agricultural products, coffee undergoes more processing by more actors along the value chain (Fig. 1) in the process of becoming a quality product (Williams et al. 2022). The making of specialty coffee is a complex process that includes both material and discursive place-making practices across spatial and power differentials to ensure material and symbolic quality attributes.

Fig. 1. The global value chain for specialty coffee in Burundi

Burundi as a coffee producer

Coffee is immensely important to the Burundian economy, as 90% of the export income is derived from coffee, and 600,000 farming families depend on coffee for their livelihoods (Lenaghan et al. 2018). The Burundian coffee sector is highly regulated by the government, and public institutions ensure that laws are followed. The regulatory authority, which approves activities before they can commence, is Autorité de Régulation de la Filière Café (ARFIC). ARFIC regulates harvests, milling, sales, and export. These highly bureaucratic procedures shift annually (L. Rosenberg 2017). InterCafé Burundi, the interprofessional association for stakeholders in Burundi involved in producing and exporting coffee, aims to improve the quality and productivity of the coffee sector. ARFIC and InterCafé Burundi collaborate with the local administration in coffee growing districts to ensure that policies are maintained. These institutions determine the operation of any value chain for coffee in Burundi. They set the coffee prices and procedures for payments to farmers, financing conditions for coffee businesses, and controlling the distribution of fertilizer, as well as the export procedures. For example, Burundian coffee farmers are only allowed to grow the Bourbon coffee varietal and to deliver their coffee to a washing station for processing, which has to export the coffee as Burundian Bourbon. There are no other known attempts to protect the use of place or to put Burundian coffee on the specialty coffee map by state agencies.

Besides the economic significance of the crop, Burundi is a unique place to study the production of specialty coffee and the making of a “taste of place.” Arabica coffee requires growing conditions with annual mean temperatures in the range 18–21°C, with no frost, but with temperature variations between day and night (Wintgens & Zamarripa 2004). Burundi has biophysical conditions with ideal altitudes and temperatures for producing specialty quality coffee (World Bank 2016; Lenaghan et al. 2018; L. Rosenberg et al. 2018). These ideal conditions are, however, challenged by climate change, which is globally predicted to reduce areas suitable for growing C. arabica coffee by 50% by 2050 (Bunn et al. 2015; Ovalle-Rivera et al. 2015). Moreover, Burundi’s challenge with climate change is the increased pest population that is directly linked to the PTD-related concerns of buyers (Jaramillo et al. 2011; Hale et al. 2022).

Burundi is known for a high frequency of PTD, which prevents Burundian coffee from qualifying for specialty coffee (Hale et al. 2022). Antestia bugs³ damage the coffee cherry skin, allowing fungi and bacteria to enter and infect the coffee bean, resulting in a foul taste (Hale et al 2022). It is the relationship between the antestia bug and the bacterial and fungal cultures specific to Burundi and Rwanda that cause the PTD. For example, Kenya and Ethiopia have antestia bugs, but they do not have the PTD, due to different microflorae (Hale et al. 2022). In a way, the PTD is a “taste of place,” but one that is not associated with quality, but rather the contrary. How, then, does the production of specialty coffee in Burundi navigate the making of material and symbolic qualities necessary for the coffee to qualify as specialty coffee? An exploration of this question is the focus for the remainder of this study. In addressing the question, this article aims to contribute to a better understanding of how the links between places, products, and quality are constructed, and how a “taste of place” is reconfigured.

Methods

This article is based on five months of ethnographic fieldwork in Burundi in 2017. Participant observation was carried out through an internship with a specialty coffee company, Mwiriwe Coffee. The company was an intermediary actor, meaning that it connected farmers with global buyers from high-value market segments of coffee (Borrella et al. 2015). Mwiriwe Coffee was started by two non-Burundian coffee professionals, when most Burundian coffee was sold as commodity coffee. The founders of Mwiriwe Coffee saw the potential to transform Burundian coffee into specialty coffee, due to ideal place-based biophysical conditions for producing specialty coffee and for increasing the well-being of coffee farmers living in poverty by linking them to the niche market with higher price points than they had received earlier. The transformation would represent a mix of business and philanthropy, putting Mwiriwe Coffee in the category of a not-only-for-profit enterprise (Höchstädter & Scheck 2015).

Both marketing and sales were done by the founders. The founders were actively engaged in establishing and maintaining long-lasting relationships with roasters/buyers, beyond a transactional business relation. The founders also invested large amounts of their time in telling the story of the Mwiriwe Coffee’s journey in Burundi. In summary, Mwiriwe Coffee was a highly personal and relational venture with the aim of putting Burundi on the specialty coffee map and alleviate poverty in the process.

The company had two washing stations (coffee processing facilities), which received and processed coffee from nearly 5000 small-scale farmers at the time when the fieldwork was done in 2017. Burundian law ensured that farmers only grew and delivered raw coffee to the washing stations, thus making the relationship between farmers and washing stations crucial for understanding coffee production (L. Rosenberg 2017).

The study targeted a particular phenomenon of interest, both in taste and climate change adaptation (Crang & Cook 2007). Accordingly, the study was a case study of Mwiriwe Coffee and of actors in an ongoing relationship with Mwiriwe Coffee.

The study was designed as an ethnographic study with participant observation due to my interest in deliberate transformation. Participant observation enabled an intimate understanding of the coffee production process in Burundi, and how that locally anchored global value chain was entangled with actors across the globe. However, the method also included the risk of bias. The risks were dealt with by using triangulation of data relating to participants, data sources, and reflexivity (England 1994; Crang & Cook 2007). The supplementary methods selected were qualitative interviews, photo-elicitation, and document analysis. The latter included analysis of internal company documents, such as quality protocols, meeting minutes, emails, newsletters from roasters, and national coffee regulating documents. In this article, the names of the company, the hills where coffee was produced, and roasters have been changed to ensure anonymity.

An overview of interviews and study participants is provided in Table 2. The interviews were conducted, recorded, and transcribed with the assistance of a translator, as many informants predominantly spoke Kirundi.

Table 2. Interviews conducted in Burundi in 2017

Position	Participants	Gender f – female m – male	Recorded interviews
Farmers	34	20 f – 14 m	22
Junior agronomists	26	7 f – 19 m	32
Staff	10	1f – 9 m	21
Founders	2	1f – 1 m	14
Hill chiefs	2	0 f – 2 m	1
Coffee buyers	12	1 f – 11 m	16
Totals	86	30 f – 56 m	106

The analysis of the data was done using NVivo 12 in three coding cycles. During the first cycle, the data were themed inductively. Based on that, the second cycle organized the data into dramaturgical coding (objectives, conflicts, tactics, attitudes, emotions, subtexts), process coding, values coding, and concept coding (Saldaña 2016). The process coding was highly relevant for this article, as I could trace coffee processing from harvest to export. The dramaturgical coding resulted in specific codes, such as “objective: produce quality coffee” and “conflict: low-quality coffee.” The most relevant code group comprised the tactic codes within dramaturgical coding, with tactics such as “teach farmers,” “catch antestia bugs,” “float coffee,” and “tell the story.” During the writing process, these codes were analyzed further, and organized into material and symbolic quality attributes.

Findings: constructing a Burundian “taste of place”

Mwiriwe Coffee established itself as a leading specialty coffee company through work on two fronts, namely quality and story, both of which were directly linked to material and symbolic attributes. They also align with material and discursive aspects of place-making. Mwiriwe started producing coffee a few years prior to 2017, when the coffee scored 84 points on the SCA scale, putting it in the “very good” specialty coffee category. However, despite qualifying as specialty coffee, it was often disqualified due to the frequency of defects.

Mwiriwe Coffee had a quality goal of 88 points, and had achieved a score of 87 points at the time when I did my fieldwork in 2017, meaning its coffee was an “Excellent” specialty coffee according to the SCA classification. Mwiriwe constructed a sense of place by actively sharing the experience of making Burundian coffee into specialty coffee on social media, and as such imbued the place of Burundian coffee production with meaning. One of the founders stated the following during a meeting with staff: “people are telling the story only because the quality is there. We need both these lenses [quality and story]. We cannot do the one without the other.” The founders believed Mwiriwe’s success depended on producing both material and symbolic quality attributes at origin simultaneously.

In the following main subsections I present three of the agencies along the value chain for coffee sold by Mwiriwe: (1) farm, (2) washing station, and (3) coffee buyers (three dark shaded boxes in Fig. 2). Material and symbolic quality attributes are presented within each agency.

Fig. 2. Studied parts (three dark shaded boxes) of the global value chain for specialty coffee in Burundi

Farm level

Material quality

Material quality attributes were produced at farm level, starting with ripe coffee cherries. The ideal sugar content in a ripe cherry in the range 18–22%. The chief agronomist at Mwiriwe stated: “It is OK from 14%, but we do not work with OK.” Practices needed to shift from a once-off harvest of all cherries to selective harvesting over the course of several months. Mwiriwe Coffee equipped farmers with laminated cards, as illustrated in Fig. 3. In addition, a range of embodied learning experiences was witnessed frequently. For instance, farmers saw the sugar content of cherries on a thermometer-like scale through a refractometer tool that staff carried with them. When tools were unavailable, I observed and agronomist handing a ripe cherry to a farmer and asking “What does this taste like?” to which the farmer responded “Sweet.” The same farmer was then given a green cherry to taste and responded that it tasted bitter and laughed. These examples illustrate how collaborative embodied learning experiences to increase an understanding of material quality attributes were aimed at ensuring quality coffee being delivered to the washing station.

Fig. 3. Laminated harvest guide given to farmers (in Kirundi, oya means no, and ego means yes)

The precondition of using ripe coffee cherries in the production process in order for the coffee to qualify as a specialty coffee was challenged by climate change, which is linked to an increase of “coffee-loving” pests (Jaramillo et al. 2009; 2011). For example antestia bugs challenge the making of a positive “taste of place.” The strategy employed by Mwiriwe Coffee was integrated pest management (IPM). The IPM strategy entailed visually locating (scouting) the insects and their eggs on the coffee trees, and then picking off and killing the insects by hand. According to one staff member, the method was a “mechanical way to work without affecting the environment.” It required a significant amount of manual labor and a commitment to producing quality coffee (M. Rosenberg 2022).

Mwiriwe Coffee hired unemployed youths from local “coffee hills” and trained them as junior agronomists as a strategy to increase the coffee quality and reduce the population of antestia bugs. Junior agronomists served as outreach agents in the “coffee hills.” They taught farmers about the sugar content in ripe cherries, how antestia bugs behaved, how they impacted coffee quality, and how they should be caught. During the daily scouting, the staff counted and recorded the number of insects caught. Shifts in farmers’ land use and harvesting practices were witnessed during fieldwork. One farmer explained how internalized the practices had become: “my friend was dressed for church but saw an antestia bug on his tree and decided to stop immediately to catch it.” Farmers were increasingly harvesting ripe cherries and catching antestia bugs.

Symbolic quality

Farmers’ stories were utilized to produce symbolic quality attributes and to build a sense of place. Mwiriwe Coffee’s story team went to the coffee growing areas and asked farmers whether they (the team members) could ask questions and take photographs to help to sell their coffee. In an informal interview setting, farmers were asked about the state of their coffee trees, their hopes, fears, and dreams. Curated portraits of farmers were captured. They were edited and published on Instagram, where most of the interaction with consumers and buyers took place. For example, a powerful head shot was accompanied by the caption: “John only knows how to read capital letters. When he was a boy, a lucky few got schooling once a week in Burundi. They only had time to learn capital letters.” Having a story team in Burundi was an active investment into humanizing Burundian “taste of place,” and it was an essential part of making symbolic quality attributes. Moreover, it was an intentional effort to make a sense of place.

Being part of the story team during fieldwork meant that I participated in curating farmers’ stories. Farmers agreed to participate in the story team’s survey, but the degree of informed and voluntary consent was at best questionable because of the unequal power distribution between coffee farmers (the providers of raw materials) and the company staff (the buyers of the raw materials). No matter how respectful and caring the interaction was, the power dimensions were unavoidable. However, I also witnessed a genuine desire from farmers to be heard and seen—something farmers said they had not previously experienced in coffee production. However, the way farmers’ stories unfolded was extractive and had hints of colonial tones to it. Having acknowledged these problematic aspects, I question whether it is possible to avoid extractive stories completely when such a power imbalance exists. The interaction illustrates the point made by critical geographers that extractive stories are intimately linked with spatial inequalities of power (Massey 1994; Harvey 1996; Cresswell 2015).

Washing station

Material quality

Farmers brought coffee cherries to the washing station to be processed by either fully washed, natural, or honey methods. Only the fully washed method is discussed here. The fully washed method was designed to maintain the material quality attributes in the coffee cherry, whilst processing the cherry into a coffee bean. Each coffee cherry went through the following stages: floating, sorting, reception, pulping, fermentation, footing, washing, soaking, pre-drying, drying, and packing.

Overall, the fully washed method was designed based on four ideas regarding material quality attributes. First, separating coffee into micro-lots by place and date of production, allowed Mwiriwe to connect the material and symbolic quality attributes. Second, the damaged coffee cherries were removed at each stage, especially those bitten by antestia bugs. This was done to avoid a high frequency of defects that disqualified the coffee from the specialty category. The third aim was to separate the coffee cherry layers (Fig. 4). The three outer layers of the coffee cherry (outer skin, pulp, and mucilage) were gradually removed with careful control to ensure that the quality of the three inner layers (the seed/bean, silver skin, and parchment) remained intact. This was based on the idea that the inner layers absorb the flavors (material quality attribute) of the outer layers, while ensuring a clean cup (material quality attribute) by removing the three outer layers that cause fermentation. Fourth, the water level within the coffee bean was controlled to avoid losing flavor (material quality attribute) by overdrying, and to avoid fungus by underdrying. The ideal moisture level of 12% was ensured with use of a moisture meter and achieved by meticulous drying procedures. The analysis showed that the essence of capturing the “taste of place” happened at the washing station by locking the material quality attributes produced at the farm inside the coffee bean, which were protected by a layer of intact parchment.

Fig. 4. The anatomy of the coffee cherry (reproduced, with permission, from Bastian et al. 2021)

Symbolic quality

Mwiriwe Coffee constructed a sense of place through insights into the daily production protocols of its washing stations in Burundi. The story team captured images of processing steps and shared stories of how quality was ensured. Stories about the material efforts ensuring material quality attributes to produce a clear representation of time and place underscore how the process is an active process of making a “taste of place,” both materially and discursively.

Producing quality coffee in Burundi came with many challenges. From my data analysis I found that Mwiriwe Coffee’s sharing of its struggles in producing coffee in Burundi built transparency and trust, which are crucial for the symbolic quality attributes and a sense of place (Daviron & Ponte 2005; Cresswell 2015). My field notes contained the following text for an Instagram post that was drafted together with one of the founders who was leading the Mwiriwe story team:

Most of our coffee production happens at night and we need fuel to power our generators so that we can have lights and run our McKinnon (pulper). We need fuel to check on the production of our washing stations. We need fuel to operate as a company. For weeks now there has been an ongoing fuel crisis in Burundi. It has forced us to bend [direct] our energy towards conserving and sourcing fuel. We now choose between spending days in petrol lines [queues] with no guarantee of fuel, indulging in the illegal market with a double price point and a risk of getting fuel mixed with water, or corrupting our way through the system.

The data collected for this study show that some challenges qualified for the public eye, while others remained backstage and never saw the spotlight in becoming symbolic quality attributes. I considered that the rationale for censoring was twofold. First, the perceived threshold of tolerance of customers and buyers was considered. The gravity of farmers’ hardships, power imbalances, price distributions, and certain quality overflows were, for instance, kept backstage. Second, many of the challenges were directly linked to the overbearing and unpredictable national governance of the coffee sector. Sharing the abundance of governance-related challenges was therefore censored to prevent putting farmers, staff, and the company at risk.

Buyers

Material quality

Buyers (roasters and importers) travelled to Burundi during harvesting, cupped the freshly processed coffee, and bought selected micro-lots on the spot. The natural limits to the amount of coffee that Burundian farmers could produce caused a limited supply of Mwiriwe Coffee. During fieldwork in 2017 I learned that roasters were becoming frustrated by coffee drinkers continually asking for Mwiriwe’s coffee, not just any Burundian specialty coffee. Inquiries to Mwiriwe Coffee from new roasters with emails with comments such as “we have a hard time selling Burundian coffee, because it is not from you” were frequent. Due to larger demand than supply, Mwiriwe Coffee let long-standing partners choose and order coffee first.

The fact that specialty coffee buyers went to Burundi to choose and order coffee confirmed (by appraising material quality attributes through cupping coffees) that farmers and Mwiriwe’s washing stations had succeeded in producing coffee classifying as specialty coffee. One buyer said he found “some incredible standouts from both Gama and Nuna [two hills], spreading everywhere from elegant citrus with sparkling acidity to deep papaya, strawberries and cream.” It is worth noting the reference to flavors and acidity, which are two material quality attributes. Another buyer found favorites from the Gama and Masa hills, and he explained that he always looked for clear flavors in the cup. He expressed excitement about the fruitiness of the coffees from the aforementioned hills that was not covered by anything else: “It was clear fruit!” A clean cup is also a vital material attribute for specialty coffee. He brought Gama coffee lot harvested in 2016 lots with him and cupped it next to a Gama coffee lot from the 2017 harvest:

It is surprisingly still amazing. The shelf life is beautiful, and that is because of the predrying and the slow drying you do. Some South American coffees just die after six months on the shelf because they dry the coffee straight on the concrete ground. The hot tar and sun bake the beans and crack the parchment. We like the parchment to be intact because the cell structure of the bean is then intact and full of flavor.

The buyer connected the material quality attributes in the cup and the shelf life of the coffee directly to the meticulous washing station procedures.

Symbolic quality

Mwiriwe sold its coffee as microlots. Each coffee bean in a microlot could be traced to a specific day and hill of production. This meant that the “taste of place” was not Burundi, but a specific hill, with its own microclimate, villages, people, and stories. Hill profiles were developed as tools to link the sense of place to the flavors experienced in the cup because of material quality attributes. The profiles were crafted by the story team based on visits to the hills.

Once a visit to a hill had been cleared with the hill chief, the story team would arrive at sunrise to capture the rhythm of life in beautiful light. Farmers were informed about the purpose of hill profiles and interviewed on the way to their fields. Details about challenges, history, culture, and the uniqueness of the hill were gathered. Later, data on location, altitude, number of farmers, coffee trees, and the flavor notes were gathered by interviewing staff. This information was edited into compelling stories in attempts to capture and convey the soul of the hill—the sense of place. The following is an excerpt from one such hill profile:

Gama hill is a stone’s throw away from the indigenous Kibira forest. The cool mist of the forest breathes daily into the coffee trees. This slightly cooler micro-climate makes the beans of Gama grow and mature slower, which sets the Gama cup apart.

As with symbolic quality attributes in general, the attributes described in the above quotation were not geographical indicators, but a sense of place and its connection to the material quality attributes.

The purpose of the hill profiles was to enable a relationship between buyers and the “taste of place.” Buyers preferred to buyg coffee from the same hills year after year, so that they could consistently offer the flavor profile that their customers liked and convey the sense of place enabled by the ongoing relationship with the hill. Roasters said it was easier to sell specialty coffee with a story because “bearded people want to know”—a reference to “hipsters” who featured in the growing demand for specialty coffee.

Discussion: taming space into a “taste of place” through material and discursive practices

How is a “taste of place” made? Producing specialty coffee in Burundi required ensuring the material quality attributes of the coffee cherry fruit. At farm level, the coffee cherry was a manifestation of the terroir (Trubek 2009; Demossier 2011; West 2022). Terroir is the foundation of the material quality of Mwiriwe Coffee, defined by the genetics of the C. arabica tree, the Bourbon cultivar, and the ideal agro-climatic conditions, which in Burundi include antestia bugs (Daviron & Ponte 2005; Hale et al. 2022). Thus, the coffee cherry is a material manifestation of relational spatial production through a sensory experience of flavor and mouth feel (Massey 2005; Trubek 2009).

The washing station was place-based and necessary to control the material quality attributes of coffee at each stage of processing (Daviron & Ponte 2005; L. Rosenberg et al. 2018). The material work of balancing the sugar content of the coffee cherries, the length of fermentation based on the temperatures at night, the harshness of the sun, and the moisture content of the coffee bean—all to make sure that the material quality attributes evolved in a controlled manner at the washing station designed to keep the material quality attributes fully intact and considered part of the terroir (L. Rosenberg et al. 2018; Williams et al. 2022).

Reconfiguring a “taste of place” in Burundi was a process of place-making achieved by the material work of ensuring material quality attributes, and the discursive work of constructing a sense of place. Prior to 2013, when Mwiriwe Coffee was founded, Burundian coffee was not only placeless commodity coffee, but also faceless. The curated stories of farmers’ profiles, washing station processes, the challenges of producing coffee in Burundi, and hill profiles were tools employed specifically to construct a sense of place with lived human experiences of place. The Mwiriwe microlots were sold with not only geographical information, but also a sense of place. Producing a place-based product that signifies quality is therefore not simply done by producing a product in a specific place (Weiss 2011). Such unregulated symbolic attributes rely on trust and transparency (Daviron & Ponte 2005; L. Rosenberg et al. 2018), which were found to exist among buyers in the study.

The growing global demand for specialty coffee plays a central role in changing how relations in coffee production in Burundi unfolded (Massey 2005; Roseberry 2005; Bro & Clay 2017). Without the demand for specific flavor profiles from the same hills year after year, it is possible that Mwiriwe Coffee would not have existed the way that it did (i.e., at the time when the fieldwork was done). Thus far in this article, the premise of social construction of fictive places has been aligned with the consumption of place-based goods (Goodman et al. 2010; Overton & Murray 2016; 2021). Furthermore, the curatorship and censorship of stories that construct a sense of place assert that a “taste of place” is a fictive place, created by actors in the value chain (Overton & Murray 2016; 2021). However, in Burundi, the efforts necessary for the coffee cherry to qualify as a quality product from a “fictive place” required more than discursive practices of consumers and experts—they also required discursive practices by producers (Weiss 2011; Overton & Murray 2016; 2021). Moreover, the material efforts of producing material quality attributes that define a “taste of place” that were essential have been grossly underacknowledged in previous studies of “taste of place.”

Fictive spaces and the role of power

It is the taming of space, materially and discursively, that makes a “taste of place.” A “taste of place” depends on the place in its material form, but it is also about the ability to tame the relations of producing space, such as global demand, climate change, local terroir, local governance, and people’s lived experiences into a place-based product deemed desirable by consumers. It is the material and discursive practices by specific relations producing space that result in both material and symbolic quality attributes (Daviron & Ponte 2005; Massey 2005; L. Rosenberg et al. 2018). The reflection on who was active in the making of “taste of place” reveals different geographies of power (Massey 1994; 2005; Cresswell 2015).

In Burundi, the state played regulatory role that also tamed relations of producing the “taste of place” in specific ways. For example, by legally detaching farmers from coffee processing at the washing stations, farmers were neither allowed to add value by processing, nor to maintain the quality of their own coffee beyond harvesting. The latter was necessary for coffee to qualify as specialty coffee, meaning that Burundian farmers’ ability to be part of the quality market segment depended on whether the washing station in their proximity focused on either specialty or commodity coffee production. Due to the large number of state-owned washing stations (Lenaghan et al. 2018), it could be questioned whether the enforcement of the law prohibiting farmers from processing coffee was a potential rent-seeking practice. Regardless of the intentions, the power that farmers had to shape the “taste of place” deliberately was highly limited.

However, the studied coffee farmers were essential for the making of Burundian “taste of place.” Without farmers’ land use and harvesting practices, Mwiriwe Coffee would simply not be possible. Without stories about farmers, such as humanizing coffee farmer John who could only read capital letters, Mwiriwe coffee would not have managed to build such a strong sense of place, with a savior touch to it. Despite the central role of coffee farmers in making both material and symbolic aspects of Burundian “taste of place,” they had very little power to impact the flows that made Mwiriwe Coffee successful in the global coffee market. Without Mwiriwe Coffee’s founders coming to Burundi from “outside’ and enabling the process of reconstructing the “taste of place” through material and discursive practices, the farmers’ coffee would not be in the global high-end niche market. Crafting a “taste of place” requires both material and symbolic quality attributes, both of which are entail a great deal of knowledge, skills, and technical tools. Coffee farmers are necessary inputs into the “taste of place” but unless they are extremely well resourced and connected, the making of “taste of place” is reserved for the privileged and connected. However, power is unequally distributed across the global political economy. For example, farmers in Burgundy, France, are able to craft their product into a “taste of place” (Hill 2020; 2022). In Burundi, both founders and farmers were essential for reconfiguring a “taste of place.” However, the different roles played by farmers, and founders in producing Burundian specialty coffee revealed unequal practices of power, and reconfirmed the spatial inequalities of power (Massey 1994; Cresswell 2015).

Conclusions

A Burundian “taste of place” was reconfigured by taming the space of coffee production into representation through material and symbolic quality attributes. This article confirms that it is possible to establish a “taste of place” in a context not known for quality. However, it also acknowledges that places with biophysical conditions viable for producing products with high material quality attributes are limited, and increasingly so with climate change.

What does it take to establish a “taste of place”? The importance of material qualities on the ground is surprisingly sidelined in studies of the discursive production of “taste of place.” The specificity of a Burundian “taste of place” included a crafted microlot of washed coffee from Gama hill, processed on a certain date. In addition, it was processed at a specific washing station with high standards for material quality set by Mwiriwe Coffee company, which worked with coffee farming communities in particular ways (e.g., by a story team, by community outreach by junior agronomists). This “taste of place” provided a differentiated experience for which buyers were willing to pay. However, the buyers were not paying a higher price for a specificity that could be captured by geographical indicators alone. They were paying for a coffee with high-quality material attributes, a trust-based relationship with the producers of the coffee, and to be part of the story of relations transforming the space of Burundian coffee production from the “ugly duckling” to a high-quality product in high demand. The study findings show that it takes both material and symbolic quality attributes to establish a “taste of place,” and that both can be altered. This article reports a study of how that can be done successfully, but it also shows how difficult and resource-intensive such work is.

A “taste of place” is not something inherent somewhere, waiting to be discovered and shared with consumers longing for quality and connection to places in a sea of placeless and tasteless products. It is romanticism of locality made possible by the unequal geographies of power. A “taste of place” is actively made by relations weaving together the worlds of production and consumption in the space of one coffee cup. It includes material and discursive practices of place-making, of taming space into representation through material and symbolic quality attributes.

Acknowledgements

The study was approved by the Norwegian Center for Research Data (on 28 June 2016, reference no. 48844). The Research Council of Norway is thanked for funding the research (grant no. 255664). Thanks are given to the two anonymous reviewers, whose comments helped me to improve this manuscript, and to colleagues Karen O’Brien, Heidi Østbø Haugen, and Yngve Heiret for valuable feedback. The many research participants are thanked for making the research possible.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1 Mwiriwe is a pseudonym, meaning “Good afternoon” in Kirundi, the main language in Burundi.

2 In the Burundian context, a hill is a geopolitical unit, known as a colline in French.

3 Genus: Antestiopis