Quantifying quality: practices of drinking water quality knowledge-making in Kaolack, Senegal

ABSTRACT

This paper examines the production of drinking water quality knowledge in Kaolack, Senegal, showing that the state apparatus of scientific knowledge production can struggle to incorporate consumer perceptions and preferences. The article builds on contributions from geography and science and technology studies which show that environmental knowledges emerge through heterogeneous social and biophysical interactions, with political consequences. Relying on eight months of ethnographic field work in Kaolack including observation, participation, and interviews, I develop a qualitative case study of four state knowledge production practices. These practices include: quantifying quality, judging sources of water, locating possible sites of knowledge-making, and attributing responsibility. Each of these practices produces boundaries around what kind of knowledge about drinking water quality is allowed to be influential and obscure elements including embodied consumer perceptions, change after delivery, heterogeneity within households and over time, and long-term effects. I argue that these processes have important implications for the inclusion and exclusion of consumers from the Senegalese state’s processes of knowledge-making about drinking water quality, with consequences for how people experience access to drinking water and exposure to potentially harmful substances in water in Kaolack.

KEYWORDS:

• Drinking water quality
• knowledge production
• Senegal
• water knowledge

Introduction

Within Senegal, the city of Kaolack is infamous for water quality issues, including high levels of fluoride in the tap water which contribute to long-term residents’ recognisability by the colour of their teeth. Still, there are ongoing disagreements between the tap water provider, who claims that the water meets required norms for potability, and consumers, who claim that the water is nearly undrinkable (MacAfee 2022). In this paper, I explore how different groups can encounter the same water in Kaolack and draw disparate conclusions based on the materiality of water quality and practices of knowledge production. I focus on how the Senegalese state produces drinking water quality knowledges in ways that make certain aspects of drinking water quality, which may be important from the consumer perspective, less visible. In particular, some relational, embodied, and emotional attributes of drinking water are relegated to a secondary role outside the realm of what is allowed to be important for drinking water quality. The inclusion or exclusion of some forms of knowledge and prioritisation of some impacts over others shapes participatory and distributive justice outcomes (Zwarteveen and Boelens 2014).

This article contributes to a broader conversation (including in this journal) about the political nature of knowledge production about environmental issues including drinking water quality, and how ongoing processes and practices around drinking water quality management normalise the unequal distribution of “drinking water” that is not always drinkable (Ayre and Mackenzie 2013; McCulligh, Arellano-García, and Casas-Beltrán 2020). Drinking water quality is largely considered to be a technical and objective phenomenon that can be known by measuring quantities of pre-determined substances in water. Current ways of measuring and responding to drinking water quality tend to assume that “ … all potential hazards are known, are measurable and have been considered” (Charles, Nowicki, and Bartram 2020, 2). However, some scholars have started to question whether standardised approaches to water quality can conceal the negotiations and decision-making behind guidelines and requirements (Bouleau and Pont 2015; Pine and Liboiron 2015; Spackman and Burlingame 2018). In particular, it has become clear that “lay” or alternative knowledges outside of the technoscientific paradigm are often marginalised (Ashdown et al. 2022), and sensory ways of knowing about water quality have been positioned as aesthetic rather than analytical and therefore also not actionable (Spackman and Burlingame 2018). Furthermore, there is growing interest in the social production of knowledge about and multiple ontologies of water (Barnes and Alatout 2012; Yates, Harris, and Wilson 2017). Thus, technocratic approaches to water quality neglect the growing consensus in the social sciences that “ … water quality is a thoroughly social category” (Linton and Budds 2014, 174). Responding to these considerations, and drawing on empirical findings and analysis from eight months of field work in Kaolack, Senegal, I demonstrate how the state makes drinking water quality knowledge into an objective-seeming and technical phenomenon. Thus the state’s practices of knowledge production tend to under-recognise the social and relational aspects of drinking water quality which may be more important to consumers.

The paper is divided into five sections. In Section 2 I present how knowledge production with regards to water has been treated in geography and social studies of science (STS). I bring these traditions together to show how knowledge emerges out of interactions between social and biophysical elements which are always situated, and I specifically highlight how these insights can and have been applied to water quality. Section 3 presents the research methodology, including a description of the study region. In Section 4, I provide a detailed empirical description of four practices in the production of drinking water quality knowledge by the state in Kaolack which create boundaries around what is able to be considered good and true knowledge of drinking water quality. These state-oriented and primarily technocratic processes are countered with aspects of drinking water quality which are perceived and understood by consumers which remain beyond the reach of the state knowledge apparatus. I conclude that drinking water quality knowledge does not exist in some kind of scientific vacuum away from social and political concerns; it is relational. Furthermore, the ways drinking water quality is made can reconfigure relations between elements and individuals, with implications for consumers’ access to drinking water and exposure to possible hazards.

Drinking water quality knowledges

The question of how and why certain kinds of environmental knowledge become dominant, and what the effects of this dominance can be, has been an important topic in a range of environmental social sciences for at least three decades (Moran and Rau 2016). The application of these concepts to water is more recent (Barnes and Alatout 2012) and thus far there have been limited attempts to specifically focus on water quality (Berry et al. 2018). In this section I briefly introduce approaches to environmental knowledge production in two fields that have pursued this issue in related but distinct ways: geography and STS. I distinguish between these fields for the sake of argument but really both have tended towards being inter- and multi-disciplinary. There is thus significant cross fertilisation between them, and the boundaries are not clear cut.

Geographical disciplines, including development studies and political ecology, have long considered environmental knowledge to be political, contested, and local. For example, studies drawing on the work of Foucault (1991; Li 2007) have shown that knowledge about the environment can be used as a technology of government by implying certain kinds of solutions and inherently attributing blame to particular subjects. There has also been an important critique of the veneration of specific types of scientific knowledge and expertise leading to exclusion of local environmental knowledge (Turnhout, Tuinstra, and Halffman 2019; Whatmore 2009). Such an approach challenges claims of objectivity and introduces the role of power, politics, and agency into analysis by beginning from a foundation that all knowledge is social as much as it is scientific. Mehta (2007, 2010), for example, showed that water scarcity discourses in Western India were constructed in ways that obscured unequal patterns of consumption and placed much of the burden on already marginalised communities. Water quality has been less thoroughly addressed, but Karpouzoglou (2012) also demonstrated that in peri-urban areas in Delhi there is an ongoing process of contestation over who is able to contribute to water quality knowledges and what kinds of knowledges are considered valid (Karpouzoglou and Zimmer 2016). More recently for studies of water there has been a turn to examining the inclusion (or exclusion) of indigenous knowledge in water management (Ayre and Mackenzie 2013; Opare 2017) and multiple ontologies and ways of knowing water (Yates, Harris, and Wilson 2017).

STS is another field that has been instrumental in interrogating how knowledge is produced and showing how all knowledge is both social and material (Latour 1993): “ … inseparable from measuring instruments, data tools, computer models and physical models, archives, databases and from all the material means we use to communicate research findings” (Srnicek 2014, 42). Starting out in the laboratory, insights from STS have been applied to understand the ways the materiality of environmental knowledge production practices and tools deeply impact on the kinds of knowledge can be produced (Lave 2015; Law 2010). As it pertains to drinking water quality, the materiality of water and substances in water and the practices of knowledge production play important roles in the kinds of knowledge that emerge and are considered viable. This interplay between materiality and knowledge extends beyond the scientific realm of experts to include how consumers sense drinking water quality with their eyes, noses, and mouths. The human sensory apparatus results in embodied experiences including taste, smell, sight, temperature, texture, and so on which are key factors informing consumer decision-making (de França Doria, Pidgeon, and Hunter 2009). Experts and scientists, on the other hand, detect drinking water quality with their tests, samples, and equipment. Human eyes, noses, and mouths are sensitive to different things than the scientific equipment the state uses to determine water quality and also produce a different kind of information. A salinometer will produce a series of numbers informing the user of conductivity indicating the salt content of water, while a human mouth and tongue will detect the presence of salt in a visceral way. Detection by human senses is not inherently less correct or accurate than that of scientific equipment but may be more difficult to communicate verbally, quantify, or compare across contexts (Spackman and Burlingame 2018).

Building on the contributions outlined in this section, in this paper I assert that drinking water quality knowledges come into being via practice as the temporary result of ongoing processes that bring into contact particular social and biophysical elements. This can be a politically inflected process which impacts on how drinking water quality is distributed, including who is exposed to potentially harmful things in their drinking water and who has consistent access to safe and sufficient drinking water.

Methodology

The empirical material presented in this paper consists of qualitative data collected during eight months in Kaolack, Senegal, from April to December 2019 as a part of the author’s PhD dissertation project. Multiple qualitative sources of data were used to develop an in-depth understanding of the phenomenon under study, namely: extended observation and participation; interviews with actors in the water sector and household interviews; and collection and review of laws, policies, and plans pertaining to drinking water quality.

The observation and participation portion of the field work lasted throughout the field work period and included households, sellers of water, and municipal staff. My time spent with households was informal and facilitated by the personal connections and language skills I developed as a US Peace Corps volunteer in Senegal from 2011 to 2013. I visited sixteen households multiple times each to socialise, discuss, and observe everyday drinking water practices. I also undertook extended periods of participant observation with actors with a stake in drinking water quality governance and provision including the Hygiene Service, the tap water provider, and sellers of drinking water. I was transparent about my role as a researcher in all my interactions, and often ended up having in-depth discussions about water quality in Kaolack wherein I shared my own perceptions and what I had learned in document reviews and via interviews in addition to asking about their perceptions and knowledge of drinking water quality. In this way, I strove to make my conversations with people into more of an opportunity for mutual learning and sharing of understanding than an extraction of information (Brinkmann & Kvale 2015).

I also carried out more formal interviews with a range of actors, including seventeen semi-structured interviews with regional and national water governance actors and eighty-four structured household interviews. Consent to participate was obtained orally following a brief description of the project, the type of data I intended to collect, and the voluntary nature of participation. Interviews were conducted in either Wolof or French. Interviews with governance actors included: planning staff at the offices of the mayor of the city of Kaolack and the prefect of the Kaolack Region; regional representatives of the ministry of environment, the ministry of urban planning, the hygiene service and the sanitation office; staff of the tap water provider; and the national department for the planning and management of water resources (DGPRE) in Dakar. I was accompanied to the household interviews by a research assistant who helped introduce me to the households and asked the questions so that I was able to observe, listen, and take notes. I selected neighbourhoods for the household interviews after conducting several of the key-informant interviews and discussing with my contacts in Kaolack about which areas of the city are most impacted by drinking water quality challenges. The general consensus was that the quality of tap water is somewhat uniform throughout the city, but that disconnected households and those further from the centre are more likely to be reliant on alternative sources with potentially lower quality. I, therefore, aimed to include a range of households in different neighbourhoods in an attempt to capture the diversity of how people actually access water and perceive drinking water quality problems in a form of theoretical snowball sampling.

In addition, I reviewed laws, policies, plans, and reports pertaining to drinking water quality in Kaolack and Senegal. In reading these documents I focused on the ways drinking water quality is defined by state actors in reference to certain kinds of scientific data and national and international standards. Then, in parallel with analysis of the interview transcripts I gradually identified patterns (in the form of similar repeated statements or groups of statements) and developed codes in an abductive process moving between the literature and empirical data (Alvesson and Sköldberg 2017).

The research project was assessed by the Norwegian Centre for Research Data (NSD) regarding the processing of personal data and deemed low-risk (Project reference number 752599). The project adheres to the NMBU Ethics Guidelines (Adopted by the University Board 26 November, 2015), and the Guidelines for Research Ethics in the Social Sciences, Humanities, Law, and Theology of the Norwegian National Research Ethics Committee (June, 2016).

Profile of the study area

Kaolack is a secondary city in Senegal with a population of approximately 300,000 people (ANSD 2019, 21). The city is an important hub for trade, transportation and agriculture, although the poorly maintained port on the Saloum River has led to a decline in influence and wealth over time. Still, Kaolack’s connection by road to every region of Senegal, Mali, the Gambia and Guinea ensures a consistent flow of traffic and trade. However, inadequate management of solid waste, particularly around the central market, and frequent flooding in the rainy season contribute to Kaolack’s notoriety as the dirtiest city in Senegal (Plan Directeur d'urbanisme de Kaolack “Horizon 2037”: Rapport Diagnostic (version provisoire), 2019, p. 240).

Kaolack is located in a hyper-saline tidally influenced estuary on the banks of the Saloum River (Dieng et al. 2017). The only available surface water, in the river, is far too salty for drinking or agricultural use, so surrounding areas are almost entirely dependent on groundwater (Dieng et al. 2017).¹ The two sources of groundwater in Kaolack are the Maastrichtian aquifer, with an average depth of about three-hundred metres, and the much shallower Continental Terminal aquifer (Diène, Kane, and Dethie 2014). Risk of salt infiltration from sea water and the saline Saloum River has led the national water authority to promote use of the deep Maastrichtian aquifer as an alternative to the Continental Terminal, but this is not without problems (Faye et al. 2005). The Maastrichtian aquifer is also slightly saline and contains significant quantities of fluoride – in some cases up to five times WHO recommendations (Faye et al. 2005). Furthermore, the tap water in Kaolack often has a briny, unpleasant taste and light brown colour (Agence Belge de Développement 2013).

Several of the substances in the tap water of Kaolack have problematic potential health impacts. For instance, salt in groundwater and the Saloum River may be associated with hypertension and pre-eclampsia (Dieng et al. 2017; Faye et al. 2005). Fluoride in the deep Maastrichtian aquifer can cause a condition called fluorosis. When exposure happens in early childhood, dental fluorosis is “characterised by discoloured, blackened, mottled, or chalky white teeth” (UNICEF 1999). Chronic intake of fluoride can also cause kidney or neurological damage and “severe and permanent bone and joint deformations of skeletal fluorosis” in both children and adults (UNICEF 1999, 11). A recent study found that 90.5% of school-aged children in Kaolack are affected by dental carries associated with fluorosis (Cisse et al. 2015). The tap water in urban Kaolack is not considered likely to have high levels of microbial contamination meaning there is limited risk of water-borne infectious diseases being transmitted by tap water consumption (UN Habitat 2009). Still, rates of diarrheal disease in children in Kaolack are nearly double the average reported among cities of similar size in Senegal (Orsola-Vidal and Yusuf 2011).

Within Senegal the Societé national des eaux sénégalaise (SONES) is the national government actor primarily responsible for ensuring the provision of safe drinking water in urban areas. In Senegal since 1995, there has been a public-private partnership for urban drinking water services consisting of a leasing (affermage) contract between SONES, the private company SDE (a division of the international company Eranove), and the state (UNDP 2011). Under the terms of this contract, SONES retains responsibility for large investments in water treatment and delivery infrastructure, while SDE maintains existing infrastructure, chlorinates the water, and provides potable water to consumers at a rate agreed upon with the government. Three independent laboratories also have contracts to conduct sporadic random sampling of the network within different regions of Senegal and report their findings directly to SONES. One of these laboratories is based in Kaolack, supported by the local subsidiary of a large international NGO. While SDE, SONES, and their independent monitors focus on tap water, the Hygiene Service under the Ministry of Health and Social Action monitors quality of other water intended for human consumption. Working with nationally established standards, regional and city level teams within the Hygiene Service test well water, bag water, bottled water, and filtered water and issue citations or revoke rights to sell in cases where standards are not met.

The standards used to determine whether drinking water is of adequate quality are based on the WHO Guidelines for Drinking Water Quality and norms established by the Senegalese Association for Normalisation (ASN).² The DGPRE works with SONES to establish policies and plans with regards to drinking water resources, such as the national IWRM action plan (Plan d’action de gestion intégrée des ressources en eau (PAGIRE)). Much of the data on drinking water quality in Kaolack represents outputs of collaborations between government actors (primarily the DGPRE or ASN) and NGOs or international organisations like the Belgian Development Agency, USAID, UN Habitat, and JICA. These include studies about solutions to drinking water quality constraints (like large scale water transfers) or various types of filtration (Lagaude, C, and Travi 1988) alongside financial analyses and feasibility studies. Epidemiological and public health studies about how compounds like fluoride and salt affect health are also common (Agence Belge de Développement 2013; DGPRE 2011). Many of these studies fall under the umbrella of the programme for potable water and sanitation in the millennium (Programme d’eau potable et d’assainissement du millénaire (PEPAM)) which was completed in 2015.

Drinking water quality knowledge production in Kaolack

In the following sections I trace the practices of knowledge production associated with the state locally in Kaolack, nationally in Senegal, and internationally in selected organisations focused on global public health. I describe empirically how the state produces drinking water quality in particular ways through their practices – particularly those pertaining to quantification, judgement of sources, location of where drinking water quality can be known, and allocation of responsibility. These practices lead to certain elements being included, excluded, or subordinated within a hierarchy of knowledge. Thus some elements appear to be important (while others do not) in terms of deciding whether or not water can or should be consumed by humans.

The Senegalese state’s ways of producing knowledge about drinking water quality are grounded in recommendations for priority compounds in water that are detailed in the WHO guidelines for Drinking Water Quality (WHO 2011, 2017a) and adapted to the Senegalese context. In line with these recommendations by the WHO, the Senegalese state works to create, control, and disseminate drinking water quality knowledge according to the norms of a certain vision of expert-produced and state-sanctioned science. These practices make certain kinds of knowledge seem valid, trustworthy, or possible, while others are dismissed or made invisible. Some potential elements of drinking water quality escape the confines of this assemblage, with particular implications for how consumers are able to access water or participate in drinking water quality assemblages. Throughout this section, I frame the discussion of these escapes around distinctions between state and consumer ways of producing drinking water quality knowledges.

Quantifying quality: tensions between acceptability and measurability

In reports and documents about drinking water quality in Senegal, as elsewhere, there is a tendency to treat quality as something which should be seen and known by experts through data in the form of measurements and comparisons of quantities of substances in water to specific standards. The data produced by these standardised measurements and comparisons are referred to more or less synonymously with properties of water. This quantitative emphasis may prevent other aspects of drinking water quality which may be important to consumers, such as taste, temperature, or relations with providers and sources, from being included in a meaningful way.

For instance, some parts of the WHO guidelines foreground scientific knowledge and devalue other ways of knowing or perceiving drinking water quality. Consider the following statement from a chapter on acceptability in the WHO guidelines for Drinking Water Quality:

To a large extent, consumers have no means of judging the safety of their drinking-water themselves, but their attitude towards their drinking-water supply and their drinking-water suppliers will be affected to a considerable extent by the aspects of water quality that they are able to perceive with their own senses. (WHO 2017a, 219; emphasis added)

The claim that consumers have no means of judging the safety of drinking water effectively diminishes the value of their knowledge and positions scientific measurement practices as the best way of determining quality. The WHO distinguishes between safety, which is knowable only through scientific means, as a privileged component of quality that consumers do not have access to, and other aspects of drinking water quality that consumers may be able to perceive with their own senses, although the guidelines do suggest that consumer acceptability is a factor that should be considered in assessing drinking water sources.

Such claims that relegate qualitative consumer knowledges to a less meaningful status are common in policies, documents and plans at the national and international level, and were echoed in interviews with national level actors in Senegal. For example, reports on drinking water in the Kaolack region assert a damaging relationship between consumers’ dissatisfaction with taste, smell, or appearance of tap water and their decision to reject “good” sources in favour of others which are more palatable but potentially less safe (DGPRE 2016; Diop 2014; Senagrosol-Consult 2009). In addition, a study of fluoride in the Kaolack region compared peoples’ stated perceptions about drinking water quality to measured quantities of fluoride and concluded: “The fact that four out of five people support that the water is good even though more than half the boreholes have a quality of water unfit for consumption due to the abnormally high concentration of fluoride bears witness to their ignorance, and above all to a misunderstanding of the risks they run” (DGPRE 2011, 25, emphasis added). This explicitly disparages the way consumers perceive drinking water quality, equating their qualitative knowledges with ignorance and misunderstandings. The implications of these reports are that consumers do not have any potentially valuable knowledge about water. Several household interview participants noted that they feel their complaints and claims about drinking water quality in Kaolack are not taken seriously by representatives of SDE, and they do not know who else they could potentially speak to in the local government that would listen to their concerns.

Of course, there are substances in water that cannot be detected by organoleptic or other means open to water consumers in an immediate sense. However, in the case of fluoride consumers are aware of the long-term connections between their water and fluorosis, or between certain sources of water and potential health impacts. Fluoride was brought up by several household interview participants in response to the question “what do you think about tap water in Kaolack?”. In response, people used visual aids like pointing to their own teeth or my research assistant’s and saying things like: “you grew up here and look at your teeth – do you have to ask if the water is good?” or “Look at my teeth!”

Health impacts from substances in drinking water are assumed to be quantifiable, pre-determined, and incorporated into standards and thresholds. However, in many cases such impacts have been abstracted to the point that they are no longer easily translatable to many people – for instance through reference to the concepts of disability adjusted life years,³ number of milligrams per litre, or tolerable burden of disease. This intangible jargon does not conjure up images of upset stomachs, discoloured teeth, or hypertension. Abstract interpretations of health impacts also obscure the differences between people, or the same people over time, and give the impression that the number of disability adjusted life years (or whichever relevant indicator) is the same for everyone. Such abstractions and technical language about water quality can limit participation of, and contestation by, non-experts (Karpouzoglou 2012). Li (2007) named such practices “rendering technical”, further emphasising the separation of environmental problems from their broader social and political contexts.

Furthermore some ways consumers know and understand quality are notoriously hard to quantify, leading these important elements to be left out, including appearance, smell, and taste. Many people also specify particular things which are sometimes present in water which give sensory information about its suitability for drinking including bleach, sand or clay, microbes, iron from old pipes, bubbles, oil, honey, sugar, salt, sediment, or soap. It is also challenging to quantify things like shame, sadness, or the degree to which being water insecure impacts other parts of your life (Wutich, Brewis, and Tsai 2020; Wutich and Ragsdale 2008). Some participants described feeling embarrassed to serve tap water to their guests from out of town, and worried about the experiences their children would have of growing up marked by fluorosis. Others were angry or disappointed because of perceived neglect by government officials who do not make Kaolack’s water struggles a priority.

Judging sources of drinking water: safely managed and improved sources

In the national and international state practices of assembling drinking water quality, another way of assessing progress towards drinking water quality goals is determination of whether each household is primarily served by a safely-managed improved source. This is proposed as a supplement to the quantitative measurement practices I described in the previous section. In cases where there are shortages of equipment, manpower or expertise for monitoring the quantities of substances in water, the type of primary source of drinking water in each household represents a secondary indicator that facilitates monitoring progress towards goals of connectivity and access.

Safely managed drinking water is defined as the use of an “improved drinking water source located on premises, available when needed, and free of faecal and priority chemical contamination”. The 2017 list of sources the WHO considers improved includes: piped household connections, public taps or standpipes, boreholes or tube wells, protected dug wells, protected springs, rainwater, tanker trucks, and bottled water (WHO 2017b). In Senegal, the national planning authority for water similarly links certain types of source to assumptions about their quality (DGPRE 2016). Unlike the WHO, the Senegalese government categorises water from wells, bags, and surface water as unclean, something consumers may turn to if they are not satisfied with the quality of tap water (DGPRE 2016). The Senegalese water law also includes standards for characteristics of individual wells to prevent possible contamination, such as distance from latrines, waste dumps, cemeteries, and so on (Faye 2017). The WHO acknowledges limitations to the improved versus not-improved indicator. They note that although improved sources are more likely to be free of contamination than unimproved, contamination is still very common.

Even so, the improved versus non-improved dichotomy situates the type of source as a more important indicator of drinking water quality than the properties of the water provided by that source (Bain et al. 2012). For example, the WHO’s 2017 drinking water quality guidelines are the first revision which accepts water delivered to the home (tanker trucks and so on) and also containerised water (sold in bottles or bags) as improved sources. The materiality of these sources did not have to change for their status to change. Instead, this change reflects shifting perceptions of those sources at the international level.

Characteristics of a source can also send signals about drinking water quality to consumers in ways that go beyond the improved versus non-improved dichotomy. Customers take note of the cleanliness of the area around a potential source of drinking water, for example, whether wells have covers, filtration kiosks are kept tidy or the containers well water is delivered in have mildew on the outside or appear dusty, indicating that the quality of water inside might be compromised. Many people I spoke to associated specific sources of water with experiences of illness or discomfort, for example claiming that filtered water or bag water gives them a cough or cold, or that water from a certain well had given them stomach problems. Concern for consistency and a keen awareness of subtle changes in the characteristics of drinking water sources are also common among consumers. Frequent cut-offs and variable taste, smell and appearance are a familiar complaint about tap water in Kaolack. Contextual elements including temperature outside and time of year can influence drinking water quality assemblages. In addition, accessibility of a source is an element of quality for many people, comprising the time, labour and financial resources required to obtain water. Accessibility is left out of the state conversation about quality and subsumed under access, although from the consumer perspective these factors influence what people decide to drink in a similar way to the quality of the water itself.

Positive or negative experiences and relationships with a particular provider of water can inspire trust or distrust in the seller or their product. Experience has, in many cases, taught consumers who to trust, not necessarily aligned with the improved sources the WHO or Senegalese government have identified. For example, some interview participants who consume filtered water regularly said that their relationship with their local provider and belief that the individuals are professional and conscientious is an important reason they believe in the quality of filtered water and feel it is safe to drink. Similarly, customers who choose to have well water delivered to their homes expressed trust in the individual who delivers to them. Those consumers who have a long-standing relationship with the delivery person believe that person is doing what is necessary to keep the water clean.

On the contrary, in the case of the tap water provider in Kaolack (Sénégalaise des Eaux [SDE]), many consumers described a lack of trust which includes suspicion about whether the company has more information than they are letting on and displeasure with their customer-client relationship. Some but not all of these complaints are directly related to the properties of the water. The chronic unwillingness of SDE to acknowledge or respond to peoples’ complaints about tap water quality arouses suspicion (Diop 2014). This lack of transparency came up even in an interview with an employee of the local urban planning office. I was told that SDE has mountains of data about water quality but it is unlikely I would ever see it.⁴ Household interview participants overwhelmingly agreed they did not have anyone in an official capacity they could ask about drinking water quality. Some also mentioned doubts about whether SDE would be willing to make their data public, or a lack of confidence in any information that SDE would provide. The combination of general feelings of distrust, frustration, and suspicion that the quality of water they are purchasing may be substandard are all elements in how consumers understand drinking water quality in Kaolack.

Locating drinking water quality

The Senegalese state also limits where drinking water quality can or should be known. For example, SDE’s contract for tap water provision only mandates measuring and monitoring drinking water quality within the confines of the piped water network, and typically with tests performed in a laboratory setting. Therefore, SDE and the NGO laboratory which monitors tap water quality follow and ensure drinking water quality until the water has crossed property lines into homes. Though they are aware of the potential for change after water leaves the tap, their obligation is limited to drinking water quality before that moment. But after water leaves the tap it often continues to move and drinking water quality can subsequently change. After waiting in a barrel or a jug for hours or even days some amount of water may be carried to a new destination for consumption or use. Households also share stores of water with their neighbours and family members during cut-offs, meaning that tap water continues to flow between homes even when pipes are empty. In addition, some companies that make bag water use tap water to fill their bags (either treated again in some way or not), and the kiosks which sell filtered water also predominantly use tap water as their raw material. Staff of SDE expressed concerns in interviews about their potential liability if someone were to get sick after consuming filtered water or bag water, and tried to distance themselves from changes in tap water quality after delivery.

Locating drinking water quality outside the boundaries of homes has additional implications for what consumers actually experience. The state and provider have obscured certain possibilities by assembling drinking water quality in ways that hide change after delivery. Furthermore, limiting the locations where drinking water quality can be known has clear effects on how the state identifies and responds to drinking water challenges. If drinking water quality is limited to the piped water network, where it is maintainable and knowable, then it is logical to improving drinking water security by connecting the maximum number of households possible. Critics of the practice of prioritising access to specific categories of safely managed sources point to a disconnect between the materiality of specific sources and their expert-assigned status. They argue that such practices incentivise continuous increases in the number of households with access to tap water because it is safely managed, without prioritising improvements in the quality of that water (Pezon 2018).⁵ Rather than focusing on access to water of sufficient quality and quantity as a unified goal, access remains independent of actual assessments of drinking water quality. Such a division allows narratives of successful increases in access to coexist with ongoing deficiencies in water quality.

The practices of measurement and assessment of drinking water quality can also obscure complexity and ongoing change. The Senegalese state mainly assesses households based on whether or not they are connected to the tap water network, with a secondary concern for other improved sources. Assessing households based on whether they are connected or not treats households as if they are internally homogeneous and primarily access water through a single source. Therefore this practice can be a stop-gap measure, but Bain et al. (2012) have shown that it tends to substantially overestimate the number of people who have access to safe water of good quality. Monitoring approaches going forward should rely on ongoing monitoring using multiple indicators for long-term security instead of one-off assessments (Charles, Nowicki, and Bartram 2020).

Household interviews and observations in Kaolack reinforced that connectivity is not synonymous with safety or long-term water security. About one third of interview participants stated that their household relies on a mix of sources of drinking water, with no consistent primary source. The majority also said that individuals and groups within the household do not all drink the same water. Elderly people in particular were more likely to avoid tap water, for reasons of health or preference. In addition, most individuals are not constantly at home and only drinking water there. So just because a household is served by an improved source does not mean everyone who lives there always drinks water that is safe.

Additionally, a household being connected to the tap water network does not mean it is forever served and that their water will always be acceptable, affordable and free from contamination. Households and individuals frequently shift between states of more or less water availability from particular sources. If a household is late on their bi-monthly payments for tap water they can be disconnected from the network. The fee to have a household’s water turned back on can be prohibitively expensive, leading to extended periods without water in the tap. Once disconnected, households that are not able to pay their bills and fees quickly enough may have their metre removed. This removal can precipitate a significant delay because there are not enough metres for every household that wants one in Kaolack, and a household that has paid off their debts can still be placed at the end of a months-long waiting list. Furthermore, tap water pressure is inconsistent and sometimes homes go for days without water in their taps. In areas where this is common households compensate by storing water in barrels (often treated with chlorine) where there is potential for contamination before the water is consumed. The physical characteristics of water and the things in and around a source of water are not frozen in time when they are measured or when water is delivered. Drinking water quality is always changing. Water flows, evaporates or stagnates; things in water dissolve, grow and settle out. Even households that are ostensibly connected to the network and subsequently considered served by an improved primary source can in practice experience a very different situation. This heterogeneity escapes the boundaries of how drinking water quality is currently assessed by the state in Senegal.

Territorialising responsibility: potability vs. quality of drinking water

The tap water provider in Kaolack also strategically territorialises what can be included in their knowledge of drinking water quality, thereby ensuring that some elements of quality are their responsibility, while others are not. To illustrate this, a representative of SDE asserted to me in an interview that the water they distribute and sell meets all important WHO guidelines, specifically referring to microbiological quality and thus potability. He acknowledged that if one considers the physio-chemical components of water this statement about tap water does not hold, but he differentiated between microbiological quality as a key component of potability and physio-chemical status of water as separate from potability and thus outside their zone of responsibility. The WHO Guidelines for Drinking Water Quality state that water should “not represent any significant risk to health over a lifetime of consumption, including different sensitivities that may occur between life stages” (WHO 2017a, 1). Both microbiological and physiochemical factors have the potential to directly cause health impacts, although the timeframe for impacts from physiochemical things in water may be relatively long-term compared to from microbiological contamination.

The somewhat arbitrary distinction between potability and quality creates a hierarchy between microbiological and physiochemical substances in water that deprioritises certain long-term health impacts. For example in Kaolack, I found that the distinction between potability and quality made by SDE facilitates claims about drinking water quality which give precedence to short-term affects from microbiological contamination and minimises the long-term consequences of exposure to fluoride and salt in drinking water. These hierarchies constructed by the state matter. If potability, framed in terms of microbiological quality, is the central element of drinking water quality, physio-chemical quality is less prioritised in policies and plans. Thus elements with the potential to impact on potential exposure of consumers to hazardous substances in their drinking water can be less visible in the state’s version of drinking water quality knowledge.

Conclusion

This paper reveals how different state practices of knowledge production about drinking water quality can make some aspects more or less visible, with consequences for residents of Kaolack. Considering the multiplicity of elements that may be part of drinking water quality knowledges demonstrates that drinking water quality is not just about objectively determining whether water is good or bad. Drinking water quality knowledge emerges contextually in a series of relations. Such political and relational aspects of drinking water quality are not often raised in discussions of drinking water quality which prioritise rational or technical knowledge above all.

Throughout the article I argue that drinking water quality knowledges produced by the state apparatus at the international, national, and local levels are delimited in ways that create gaps between how the state and consumers understand drinking water quality. These gaps can allow some things to escape or become invisible in certain knowledge regimes around drinking water quality, with implications for access and exposure (and exclusion of non-state knowledges). Many elements can escape the state’s rigid, bounded understanding of drinking water quality. These escapes are made possible in some sense by the limits the state places on what can be known about drinking water quality, who can know about it, and how and where it can be known. Although drinking water quality includes many elements which allude to the multifunctional nature of quality, these elements are arranged into a hierarchy where some are more important than others. These practices also obscure the normative and values-based nature of “quality” as an idea and hide the negotiations and political decision-making inherent in any definition of water quality (See for example Bouleau and Pont 2015).

The arrangement and prioritisation of elements in the drinking water quality knowledge practices presented in this article can also have implications for recognition and participation of consumers, both of which are integral parts of a water justice framework (Zwarteveen and Boelens 2014). Both devaluing of consumer knowledge and rendering technical through abstract notions of drinking water quality circumscribe the ability of consumers to make claims about drinking water quality or contest unsatisfactory conditions in ways the state would recognise (Karpouzoglou 2012). Instead of trying to understand and communicate across difference, the state generally denies existence or value of other ways of knowing water quality and says theirs is the only one that matters or should influence decision-making. When it comes to distinguishing between consumer and state knowledges of drinking water quality, the point is not to reject scientific knowledge in favour of consumer perceptions, but to make that knowledge transparent and available to consumers while also taking into account their concerns about acceptability, affordability, and so on. Preventing expert knowledge produced by state actors from being available to consumers can make it more difficult for them to make informed decisions about their drinking water, although, as described previously, consumers may not trust information from drinking water providers anyway.

Furthermore, some of the elements the state de-prioritises or excludes can make people more vulnerable to negative impacts from water quality hazards. Actors associated with the state and the tap water network primarily measure water in pipes and storage tanks, which means they may end up missing changes in quality after delivery or where people actually consume water. The state practice of categorising households as either served or not served by an improved source can also obscure heterogeneity of user practices inside homes, along with inconsistency of access and quality of each source. This approach can obscure differences within and between households, hiding the fact that even in homes with piped network connections, people are not equally impacted by water scarcity, affordability, or exposure to unsafe water. It can also make change after tap water delivery invisible to the state apparatus. Such blind spots allow the Senegalese state to continue to prioritise connectivity over quality in their efforts to improve water access, leading to ongoing delivery of hazards into homes (Senagrosol-Consult 2009). They further make groups with less financial or social capital within households more vulnerable to exposure to harmful things in drinking water. In addition, the hierarchies constructed by the state matter. If potability, framed in terms of microbiological quality, is the central element of drinking water quality, physio-chemical quality is less prioritised in policies and plans. Thus elements with the potential to impact on exposure of consumers to hazardous substances in their drinking water can be invisible or escape the state’s practices of producing drinking water quality knowledge. All of these blind spots and escapes obscure the nuance and complexity in how water actually arrives to people where they consume it, and the diversity of drinking water quality in space and time.

In this article I have shown that each type of knowledge about drinking water quality in Kaolack contains provisional collections of distinct elements, and also ascribes different weight and importance to those elements in their processes of knowledge-making. The focus on elements and practices emphasises that drinking water quality knowledges are made (and therefore could be made in different ways). By considering the ways current knowledge-making practices exclude or territorialise what kinds of knowledge are possible, it may become possible to see new possibilities to envision more just, equitable, or inclusive drinking water quality knowledges in the future.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 It is challenging to lay out the background of water quality in the Kaolack region without giving some kind of privileged position to the way it has been presented in academic and grey literature. Although there may be limitations to the studies that produced this knowledge, I feel it is still useful to give a description of the hydrological context, in full recognition of the fact that this information may be contested. The challenge is to “take due notice of natural scientific accounts of the specificity of particular materials, while not assuming that this specificity should be understood only in natural scientific terms” (Barry 2013).

2 The primary law for water quality is the Loi Portant le Code de L’eau (Loi n. 81-13, 1981) and national norms for drinking water quality are detailed in NS 05-033-1996: Norme sénégalaise sur la qualité des eaux d’alimentation humaine.

3 The WHO uses a threshold of 10⁻⁶ disability adjusted life years as an acceptable level of risk for each regulated compound.

4 This was an accurate prediction. After several visits, conversations, and follow-up text messages it became clear that SDE did not want to explicitly tell me no but did not intend to share their data with me. This may also have been related to the fact that SDE was engaged in legal action regarding the termination of their contract with SONES during the period of my field work.

5 Similar trends have been identified in South Africa (Rodina and Harris 2016) and globally (Shaheed et al. 2014).