Feeling Rules and Domestic Emotion Work Under Chronic Water Scarcity

Abstract

What happens to emotional life when running water fails not for a day but for a decade? Drawing on 20 semi-structured interviews in a rural Romanian commune that went more than ten years without a functioning public water supply, and where a new network now reaches only part of the settlement, we examine how the emotional demands of chronic water scarcity are socially organised. We identify four recurrent patterns, which we analyse as feeling rules: shame before visiting outsiders, worry tied to falling household reserves, hope maintained in public, and resignation expressed in contained, formulaic terms. The four carry unequal normative warrant, which we assess explicitly. We propose the concept of infrastructural emotion work, the unpaid emotion work generated by the sustained failure of systems socially expected to function, and identify formulaic emotional display as one of its mechanisms, in which a conventional phrase provides an interactionally acceptable response while foreclosing elaboration. We argue that these expectations are socio-materially generated rather than culturally prescribed alone, and that the resulting burdens are unevenly distributed by gender. Where infrastructure is unreliable, households assemble an emotional infrastructure of their own.

1. Introduction

Running water is, among other things, an emotional infrastructure. When it works, nobody thinks about the feelings involved in flushing a toilet or offering a guest a glass of water. When it fails, a set of emotional demands appears that was previously unnoticed. The host must explain, the guest must pretend not to mind, someone must decide whether today is the day to do laundry or whether the reserves should be saved for drinking, and someone must absorb the frustration when the answer is no. This is the pattern Star (1999) described for infrastructure generally: it is invisible in operation and becomes an object of attention only in breakdown. What breakdown makes visible here is not only pipes but a set of socially organised responses that follow identifiable expectations, require sustained effort, and fall unevenly on different household members. They are, in short, emotion work, and they are produced not by an employer or a market but by the absence of a pipe.

This article examines the emotion work that households perform when water scarcity becomes a permanent condition of domestic life. The setting is a rural commune in eastern Romania, which we call by the fictionalised name Valea Secată, that went more than a decade without a functioning public water supply. The commune is not geographically isolated: it lies near a small town where water flows normally, in a county that is not among the poorest in the country, and within a European Union member state. In the year before our fieldwork, a regional network was finally commissioned along the main road, leaving households on side lanes unconnected, so that the commune now contains both connected and unconnected streets. The absence of running water is therefore not a development failure in a remote region but an infrastructure gap sustained by a specific history of institutional neglect, documented in a previous publication (Cocîiu and Rughiniș 2025). That earlier article described the ecological crisis, the economic costs, and the phenomenon of climate silence in the community. The present article shifts attention from collective discourse to household-level emotional management.

A growing body of research documents that water insecurity produces emotional distress: anxiety, anger, shame, depression, and interpersonal conflict (Wutich 2009; Wutich et al. 2020; Kimutai et al. 2023). This literature treats distress primarily as an outcome of scarcity, a dependent variable produced by inadequate supply. What it does not examine is how that distress is socially organised: who is expected to feel what and before whom, and who does the work of managing those feelings. The shift from outcome to managed process is the analytical move this article proposes.

The framework for this shift comes from Hochschild’s (1979, 1983) sociology of emotions. Hochschild introduced the concepts of feeling rules, the social norms that prescribe what one should feel, display, or suppress in a given situation, and emotion work, the effort individuals make to bring their feelings and their expression of them into alignment with those norms. She developed these concepts in the context of paid service work and later extended them to the gendered politics of domestic labour (Hochschild 1989). We apply them to a context she did not consider: domestic life under chronic infrastructural scarcity, where the demands on feeling are generated not by an employer or by cultural convention alone but by the material conditions of daily provisioning.

Our research question is accordingly: how do residents of a commune experiencing chronic water scarcity describe, evaluate and manage the emotional demands it creates, and what shared expectations organise those processes?

Drawing on 20 semi-structured interviews, we identify four recurrent patterns of emotional expectation: a shame rule governing encounters with outsiders who come from places where water flows; a worry rule governing responses to falling household reserves; a hope rule requiring public maintenance of optimism about the future; and a resignation rule permitting expressions of defeat only in contained, formulaic terms. We show that these expectations are unevenly distributed, with women carrying a disproportionate share of both practical water management and the emotional maintenance of household functioning, and we are explicit about the fact that the four patterns are not equally well evidenced as feeling rules.

The article develops one principal concept, infrastructural emotion work, to name the unpaid emotion work produced by the sustained failure of material systems that are socially expected to function. Three arguments elaborate it: that the feeling rules at issue are socio-materially generated rather than culturally prescribed alone; that the emotion work they demand is temporally differentiated rather than uniform; and that one of its characteristic mechanisms is what we call formulaic emotional display, in which a conventional phrase provides an interactionally recognisable response while foreclosing elaboration. A fourth finding, that the resulting burden falls unevenly by gender, we treat as an empirical result rather than as a conceptual contribution.

2. Literature Review

2.1. Emotion Work, Feeling Rules, and Domestic Life

Hochschild introduced emotion work and feeling rules to describe how individuals manage their inner states and outward displays to align with socially shared expectations about what should be felt in given situations (Hochschild 1983). In The Managed Heart, she distinguished emotion work, which occurs in the private sphere under informal social norms, from emotional labour, performed for a wage under an employer’s direction. Both involve two principal techniques: surface acting, in which the display is adjusted without changing the underlying feeling, and deep acting, in which the individual works on the feeling itself. The norms specifying which emotions are appropriate, how intensely, and for how long, she termed feeling rules (Hochschild 1979). Because this study concerns unpaid domestic settings we use emotion work throughout, reserving emotional labour for the paid contexts in which Hochschild coined it. Jderu (2012) provides a synthesis of the field in Romanian sociology.

The Second Shift extended the analysis into domestic life, documenting the unequal distribution of household labour between husbands and wives, including its emotional dimension (Hochschild 1989). Women performed a disproportionate share of the work required to maintain the household’s emotional equilibrium: soothing tensions, managing children’s feelings, absorbing frustration, sustaining a positive atmosphere. Tereskinas (2022), studying Lithuanian middle-class fathers, found that while men increasingly claim emotional competence in childrearing, the daily practice of emotion work remains structured by gender, with mothers performing more sustained and less visible maintenance. The concept of emotional capital, which he uses following Nowotny (1981) and Reay (2000), refers to emotion-based knowledge, management abilities and capacities to feel that can be activated as a resource within family life. Grünberg and Matei (2020) argue that the work–family conflict paradigm, by treating vulnerability as constitutive of women’s position, obscures the social imaginaries through which women themselves make sense of that work.

The framework’s applicability beyond the workplace has been demonstrated in caregiving. Halevi Hochwald et al. (2022), studying family caregivers of people with end-stage dementia in Israel, showed that feeling rules in caregiving include putting the care-recipient’s needs first, protecting the person, and avoiding conflict, and that the dissonance between what caregivers actually feel and what these rules prescribe is a source of burden but also, paradoxically, a coping resource. This dual character of emotion work, as both burden and resource, recurs in our material.

Goffman’s work on the presentation of self and interaction ritual provides a complementary framework for analysing how individuals manage face in encounters with others (Goffman 1959, 1967). Face-work refers to the actions taken to ensure that the impression given in an encounter is consistent with a claimed social identity; when the conditions of interaction threaten face, corrective processes follow, including avoidance, concealment and ritual repair. Jderu (2010) develops the dramaturgical account of emotion in the same tradition. Lavie and Mayer (2025), studying food and beverage workers during the COVID-19 pandemic, confirmed that the demands of emotional management intensify when the conditions of interaction are disrupted by forces beyond the actor’s control. The theoretical point transfers: when material conditions destabilise routine interaction, the demands on emotional management increase.

Relevant to the mechanism we develop below, Hebson et al. (2007) used the phrase formulaic emotion work in their study of English teachers subject to capability procedures, describing a shift from what they called philanthropic emotion-giving to prescriptive and standardised forms of emotional conduct demanded by a performance regime. Their construct is institutionally prescribed: an employer specifies the form that emotional conduct must take. The mechanism we develop in Section 5.2 is related but distinct, since no institution prescribes anything in our setting.

2.2. Shame, Dignity, and the Emotional Life of Poverty

Sociological research on poverty has increasingly attended to its emotional dimensions, and shame occupies a central position. Sayer (2005) argued that people care about their class position not only because of its material consequences but because class affects their capacity to live in ways they consider worthy of recognition. He distinguished conditional recognition, which depends on what people do and achieve, from unconditional recognition, owed to persons simply as persons; class inequalities undermine the former by restricting access to socially valued practices and goods. The moral sentiments associated with class, including shame, embarrassment, pride and resentment, function as evaluative judgements about how social arrangements affect well-being (Sayer 2002, 2005).

Chase and Walker (2013) examined how shame operates in the lives of people experiencing poverty in the United Kingdom. Drawing on 42 in-depth interviews, they found that shame in poverty is co-constructed, arising both from internal self-evaluation and from external interactions in which people are made to feel diminished. Participants rarely used the word shame directly, preferring colloquialisms such as awkward, embarrassed, guilty, degraded and useless, what the authors call the colloquialisms of shame. They also showed that people living in poverty develop strategies to manage shame: withdrawal from interactions that might expose hardship, pretending to cope, and constructing hierarchical distinctions between themselves and others who are worse off. This last strategy protects individual dignity but fragments solidarity among those who share the same conditions.

Stambe and Parsell (2024) argued that shame is overemphasised as the dominant feeling of poverty. Their ethnographic research across two Australian charity centres found hope and anger sitting alongside shame: service providers cultivated hope for a better future, while people in poverty held smaller versions of hope informed by everyday struggle, and were also angry at hope lost. Their concept of a political economy of hope, in which hope is unevenly distributed and can function as a placeholder that fails to disrupt the conditions producing hardship, is directly relevant here, where shame, worry, hope and resignation are co-present and mutually shaping rather than discrete. Ferreira et al. (2023), studying professionals working with domestic violence victims in Portugal, similarly found that coping strategies can both mitigate and reproduce emotional strain.

2.3. Water Insecurity and Emotional Distress: From Outcome to Process

A substantial literature links water insecurity to adverse mental health outcomes. Kimutai et al. (2023), in a systematic review and meta-analysis of 26 studies, found that water insecurity is associated with increased odds of common mental disorders, with a pooled odds ratio of 2.05 and a stronger association among women. Wutich (2009), in an early study of a Bolivian squatter settlement, documented that women reported greater emotional distress related to water insecurity than men, attributing the disparity to their primary responsibility for household water management. Wutich et al. (2020) confirmed the association between water insecurity and distress, worry, anger and helplessness across diverse settings.

Qualitative research has begun to specify the mechanisms involved. Achore et al. (2020), synthesising 21 studies, identified nine household coping strategies, several with adverse psychosocial consequences, and noted that reliance on labour-intensive strategies deepened economic disadvantage. Bose et al. (2024), studying rural Gambia, found that reductions in hygiene, changes to food consumption and long-term storage were common, and that social factors including face-saving motivations influenced which activities were maintained. Collins et al. (2024) argued that existing research documents what people do but rarely examines the emotional and social costs of doing it.

The gendered dimensions are well documented. Abuzerr et al. (2025), studying internally displaced women and girls in Gaza, found that water, sanitation and hygiene insecurity disproportionately affected women through health risks, menstrual hygiene challenges and psychosocial distress. Subah et al. (2024), examining urban flooding in Dhaka, documented that women bore the burden of both reproductive and productive work while their control over resources declined sharply.

Studies of water governance supply a further context. Sondershaus and Moss (2014) showed how, absent effective formal institutions, stakeholders’ constructions of vulnerability translated into conflicting resilience strategies, so that one actor’s resilience increased another’s vulnerability; their notion of institutional voids resonates with our case. Gill et al. (2026) applied Nixon’s (2011) slow violence to the chronic water crisis behind the 2022 Pearl River flood in Jackson, Mississippi, finding resource losses, health concerns and low trust in government associated with elevated psychosocial stress. Chelcea’s work on water in New York restaurants and households (Chelcea 2023, 2024, 2026) attends to how the material conditions of access shape social relations even where water is present but variously trusted. For the Romanian rural context, Dorondel and Posner (2022) examined how broken socialist-era agricultural infrastructure reshaped groundwater patterns and local power relations in a Lower Danube village, a structural parallel to ours. Chelcea (2002) documented household provisioning under the chronic shortages of 1980s rural Romania, a repertoire that the present case in part revives.

Despite the volume of this research, a consistent pattern emerges: distress is treated primarily as an outcome of exposure rather than as a process households actively manage.

2.4. Hope, Resignation, and Cruel Optimism Under Chronic Hardship

The sociology of poverty has given increasing attention to the temporal and affective dimensions of chronic deprivation. Edmiston et al. (2025), drawing on qualitative longitudinal research with people on the lowest incomes in the United Kingdom, introduced the concept of deep poverty to capture forms of privation falling outside conventional sociological attention, documenting how successive crises and welfare retrenchment eroded survival strategies, social connections and health. A methodologically important finding was that those experiencing the deepest poverty were also most likely to fall outside the research process itself, which bears on our own use of insider access.

England (2025) applied Berlant’s concept of cruel optimism to frontline homelessness work in Wales, describing a relation in which attachment to a desired object is itself an obstacle to flourishing. Workers systematically cultivated optimism in clients as a mechanism for survival within a hostile housing system, perceiving this as both a moral and a pragmatic necessity even as it sustained compliance with a system that offered little. Hope, on this account, can function as a governance tool, and its cultivation involves specific emotion work performed by identifiable actors. Das (2015) argued that sustaining daily existence amid institutional breakdown requires ongoing moral and emotional work; Hamilton (2022) described how climate activists transform sorrow into determination through collective emotional processes. The shared question is how people sustain emotional functionality when their conditions offer limited grounds for the positive affects that social participation ordinarily presupposes.

In the Valea Secată case, hope and resignation are not opposed states but coexisting orientations. Formulaic expressions circulate for both, and, as Section 4.3 shows, they are used in ways that contain rather than release the frustration they name.

2.5. Infrastructure, Breakdown, and Affect

Because our argument concerns emotional demands generated by a failing material system, it also draws on the study of infrastructure. Star (1999) established the analytical point on which our account rests: infrastructure is characterised by its invisibility in normal operation and becomes visible, and an object of explicit attention and work, upon breakdown. Anand (2017), studying water in Mumbai, showed that access to a public network is not a binary condition but an incremental, intermittent and reversible relationship, which he termed hydraulic citizenship, in which recognition by state agencies determines who receives legitimate service. That formulation describes our setting closely, since the commune now contains connected main roads and unconnected side lanes, and since connection has not resolved concerns about the water that arrives. Rodgers and O’Neill (2012) proposed the concept of infrastructural violence to capture how marginalisation and disconnection become operational and durable through infrastructure, which complements Nixon’s (2011) slow violence by locating the harm in specific material arrangements. Postsocialist infrastructure research shows connection distributed unevenly within settlements (Chelcea and Pulay 2015) and disconnection enacted through identifiable institutional events (Zamfirescu and Chelcea 2021).

Work at the intersection of infrastructure and affect has developed rapidly. Bosworth (2023), reviewing the notion of affective infrastructure, distinguishes two established senses: infrastructure understood as something that generates emotions, and affect understood as what organises political action. Carr et al. (2026), studying Australian community centres, describe an infrastructural politics of emotion in which staff and users anticipate discomfort, manage expectations and develop routines within the infrastructure they inhabit. Our concept is adjacent to both but distinct. It concerns neither the emotional effects of infrastructure as such nor the affective organisation of politics, but the unpaid, distributed and largely unrecognised work of managing feeling that arises when an expected system does not function, and the socially shared expectations that organise that work.

2.6. From Organised Denial to Household Emotion Work

Two conceptual lenses connect collective emotional arrangements to ecological crisis. The first is organised denial, developed by Norgaard (2011), who studied how residents of a Norwegian town collectively avoided engaging with climate change despite direct evidence of its effects, and argued that this silence was not ignorance but an active social achievement maintained through conversational norms, selective attention and emotional management. The second is the sociology of climate emotions, developed by Neckel and Hasenfratz (2021), who distinguished publicly visible climate emotions from the less visible dynamics of emotional suppression and affective concealment operating beneath public discourse, and showed that the emotionalisation of ecological crisis is characterised by a high degree of reflexivity. Vulpe (2024), analysing Romanian online discourse, shows how climate-sceptical positions are constructed argumentatively, which is the discursive counterpart of the silence Norgaard describes.

In a previous article (Cocîiu and Rughiniș 2025) we documented climate silence in this community, showing how residents collectively avoided framing their water crisis in terms of climate change despite living with its consequences. The present article moves from collective denial to household-level emotional management. Organised denial helps explain why the water crisis is not publicly named as a climate problem, while emotion work explains who bears the daily cost of living within that silence.

2.7. Gaps and the Present Study’s Contribution

Three gaps follow. The literature on water and mental health treats distress as an outcome rather than a managed process, leaving unexamined the expectations governing what may be felt and displayed and the allocation of the work of managing them. Hochschild’s framework has been applied to service work, caregiving and family life, but rarely to domestic life under chronic infrastructural scarcity, where the demands on feeling are generated neither by an employer nor by cultural convention alone. And work on infrastructure and affect has examined the emotional effects of infrastructure and the affective organisation of politics, but not the household labour that chronic infrastructural failure requires.

3. Materials and Methods

3.1. Research Setting

The study was conducted in a rural commune in eastern Romania, in a hilly area at the foothills of the Subcarpathians. The commune comprises several villages and went more than a decade without a functioning public water supply network. Its water history has three phases relevant to this analysis. A rudimentary network dating from the socialist period, fed by forest springs and public fountains, fell into disuse; households then depended on private wells, many of which diminished or dried up under recurrent drought, on springs at some distance from homes, on purchase from vendors and shops, and on cistern deliveries organised by the commune. Then, in mid-2025, a regional water project that had been discussed and repeatedly delayed was commissioned along the main road. Because our interviewing was conducted in two periods, in June and July 2025 and again in April and May 2026, the study spans that change: one respondent interviewed in July 2025 had been connected three weeks earlier, another was connected in the autumn of 2025, and by spring 2026 respondents estimated coverage at between about a quarter and about half of households, with those on side lanes still unconnected. Several respondents drew attention to the resulting inequality within the settlement. Connection also did not settle the matter of water quality: respondents in connected households reported that the network water smelled strongly of chlorine and continued to buy bottled water for drinking and cooking.

The commune lies near a small town where water supply functions normally, and this proximity makes the absence of running water a daily point of comparison. The ecological and institutional dimensions of the crisis are documented in a previous publication (Cocîiu and Rughiniș 2025), which described patterns of drought, infrastructure failure, economic costs and climate silence. The present article shifts attention from collective discourse about the crisis to the household-level management of the emotions it produces.

3.2. Sample and Recruitment

Twenty semi-structured interviews were conducted with adult residents of the commune, in two periods: five between 24 June and 8 July 2025, and fifteen between 23 April and 3 May 2026. Respondents were recruited through a combination of the first author’s personal contacts in the community and snowball sampling. The sample includes women and men from four age bands, with varied occupational profiles spanning education, healthcare, public administration, private-sector and service employment, retirement, and unpaid domestic work. Household sizes ranged from two to six members, and duration of residence from a few years to a lifetime. Because recruitment proceeded partly through personal contacts, two respondents belong to the same household; the sample therefore comprises 20 respondents from 19 households. This was not planned, but it permits a limited comparison of how members of one household describe the same conditions from different positions in its division of labour, and we use it for that purpose in Section 4.4.

The sample also spans the uneven connection described in Section 3.1. Four respondents, in three households, had network access at the time of interview, one of them for only three weeks; the remainder did not. One respondent had left the commune for a nearby town because of the water situation. The sample was not designed to be statistically representative but to capture a range of household situations, age groups, and relationships to the water crisis within a single commune.

Table 1. Respondent profiles (N = 20).

Interview	Age Band	Gender	Occupation Group
Int. 1	18–26	M	Service sector
Int. 2	27–36	M	Education
Int. 3	18–26	F	Education
Int. 4	27–36	F	Unpaid domestic work
Int. 5	27–36	F	Unpaid domestic work
Int. 6	37–55	M	Private sector
Int. 7	Over 55	F	Retired
Int. 8	Over 55	M	Retired
Int. 9	27–36	M	Private sector
Int. 10	27–36	F	Healthcare
Int. 11	37–55	F	Unpaid domestic work
Int. 12	37–55	M	Service sector
Int. 13	Over 55	F	Healthcare
Int. 14	Over 55	M	Retired
Int. 15	Over 55	F	Retired
Int. 16	18–26	M	Private sector
Int. 17	27–36	M	Private sector
Int. 18	18–26	F	Public sector
Int. 19	37–55	F	Education
Int. 20	37–55	F	Public sector

Note: Age bands are 18–26, 27–36, 37–55, and over 55. Occupation groups aggregate specific roles to prevent identification. Respondents are identified in the text by interview number only. Village identifiers are omitted to prevent triangulation, and duration of residence is not reported by respondent because, in combination with the other attributes, it would render individual rows identifiable. Sample composition: 11 women, 9 men; 4 respondents aged 18–26, 6 aged 27–36, 5 aged 37–55, and 5 aged over 55.

presents the respondent profiles.

3.3. Researcher Positionality and Insider Access

The first author is a native of the commune who has lived with the water situation throughout his life and is a doctoral student in sociology. The other three authors are not members of the community and contributed external methodological, resourcing, and supervisory perspectives; the second author took the primary role in transcript-based interpretive dialogue. This section describes how the insider and external positions shaped the analysis. The study is not autoethnographic: it includes no fieldnotes, personal narrative, or systematic analysis of the first author’s own experience as data. What insider status provided was access and interpretive knowledge.

Respondents spoke with the first author as a neighbour and co-resident rather than as an external interviewer, which allowed conversations to reach emotionally charged material with less interactional friction and made available the community’s conversational norms, formulaic expressions and shared references. Edmiston et al. (2025) note that those living in the deepest deprivation are also most likely to fall outside research altogether; insider access is one means of reducing that exclusion.

It also carried risks, and not only the obvious one. Familiarity with local expressions made it easy to treat a formulaic answer as self-evidently meaningful rather than as something requiring analysis, and shared experience of the crisis created a standing temptation to read agreement into ambiguous responses. The second author’s distance was useful against both, but external scrutiny is not simply a corrective applied to insider bias, since distance carries its own liability: an outside reader is prone to over-interpret unfamiliar idiom and to miss when a phrase is doing conversational rather than expressive work. We therefore treated disagreement between us as informative rather than as error to be resolved by seniority, and where our readings diverged we returned to the Romanian transcript and asked which required fewer assumptions about what respondents had not said. This changed the analysis in at least one consequential way, described in Section 5.2: an interpretation that a formulaic answer indicated the absence of underlying feeling was rejected as unsupportable from interview data, and the claim was reformulated in terms of what the transcripts show.

3.4. Interview Procedure and Ethics

Interviews were semi-structured and followed a topic guide covering daily water practices, emotional responses to scarcity, household decision-making about water use, interactions with outsiders and guests, perceptions of institutional responsibility, and views of the future. The guide asked respondents not only what they do about water but how they feel about what they do. Interviews were conducted in Romanian and audio-recorded with consent. The fifteen interviews of 2026 followed the guide described above; the five conducted in 2025 used a related guide developed for the study reported in Cocîiu and Rughiniș (2025), which covered drought, water access and climate perceptions but did not include the questions on gendered impact. Section 4.4 reports the resulting denominators rather than treating the sample as uniform on this point. Most interviews were conducted with one respondent alone; in at least two a family member was present and occasionally contributed. Only the designated respondent’s speech was coded and quoted; the family member’s presence was considered solely as interactional context, since what can be said about a spouse’s feelings changes when the spouse is in the room.

Interviews lasted between 11 and 52 min, with a median of 17; six ran under 15 min and five over 25. Given that range, the relationship between length and analytical value deserves comment. The shortest interviews were not the least informative and were in fact disproportionately important here: the shortest of all supplied the account of shame in a newly built house in Section 4.1, and two others under 15 min supplied the formula of acquiescence in Section 4.3 and the proverb analysed in Section 5.2. This is partly because brevity is itself among our objects of analysis: where a respondent answers a question about the future in four words and offers nothing further, the shortness is the finding, and interviews composed largely of such answers are correspondingly short. What brevity does limit is triangulation within a single interview, so we rely on patterns recurring across interviews rather than on any single account, and treat the two consistently brief interviews as deviant cases to be analysed rather than as thin data.

Ethical safeguards were integrated into the design. All participants were adults and participation was voluntary. At the start of each interview participants were informed of the purpose of the research, of the confidential and anonymised handling of the data, of its exclusive use for doctoral scientific research, of the voluntary nature of participation, and of their right to stop at any point; agreement to audio recording was requested and recorded before the interview proceeded. Because the study addressed domestic experiences involving shame, worry, hope and resignation, the interviewer did not press respondents beyond what they wished to disclose. The commune is identified by a fictionalised name, village identifiers are omitted, respondents are identified by interview number only, and age, occupation and residence categories are aggregated or withheld to reduce the risk of identification. Details of former public office, occupational specifics, and household composition have been generalised in reporting for the same reason.

3.5. Transcription, Translation, and Analysis

Because the analysis turns on the interpretation of particular Romanian words, proverbs, hesitations and qualifiers, the handling of language requires description. Interviews were transcribed in full in Romanian by the first author, a native speaker, and all coding and analysis were carried out on the Romanian transcripts; no analysis was conducted on translated text. Quotations were translated into English by the first author only at the point of writing, and checked against the original by the second author, who also reads Romanian. Translations are close rather than literal: hesitations, repetitions, unfinished sentences and colloquial register are preserved where they carry analytical weight, because in several cases the hesitation is the finding. Where a Romanian term has no close English equivalent, or where our reading depends on the specific word chosen, we give the Romanian in square brackets and explain its semantic range rather than settling it silently in translation. The risks here are those identified by Temple and Young (2004) and by van Nes et al. (2010), who address the particular case in which researchers and participants share a non-English language and publication occurs in English: translation is itself an interpretive act, and moving too early from the source language can foreclose readings the original leaves open. Our protocol addresses this by keeping analysis in Romanian and translating last. Supplementary Material provides the Romanian source passages and English translations underlying every interview quotation, together with a small number of passages closely paraphrased in the analysis.

Analysis followed a theoretically informed thematic analysis combining deductive sensitisation with inductive identification of patterns, in the hybrid form described by Fereday and Muir-Cochrane (2006). The deductive element was Hochschild’s distinction between feeling rules and emotion work, which directed attention to passages where respondents described emotional states, the management of those states, or expectations governing their expression. The inductive element was the identification of the four specific clusters reported here, which were not specified in advance and did not correspond to the topic guide’s sections. Coding proceeded in three rounds: first, identifying all passages concerning emotion in relation to water; second, grouping these into candidate clusters and testing whether each cluster described a socially organised expectation or only a recurring emotion; third, examining how the resulting work was distributed within households, with attention to gender. Deviant cases, where respondents departed from the dominant pattern, were treated as analytically productive and are reported in Section 4.

The judgement that twenty interviews were sufficient rests on the aim rather than on a numerical rule. The study seeks to identify shared expectations governing emotional conduct and to specify who performs the resulting work; it does not estimate how prevalent or intense any emotional response is in the population, and no such claim is made anywhere in what follows. For the narrower aim, adequacy depends on whether the sample spans the positions from which those expectations might differ. Ours spans both genders, four age bands, six occupational groupings, household sizes from two to six, residence from a few years to a lifetime, connected and unconnected streets, and one household that had left the commune altogether. By the final interviews, accounts of the four clusters were recurring without new elements, while accounts of connection and water quality were still varying, which is one reason we treat the consequences of connection as an open question in Section 5.4 rather than as a finding.

3.6. Identifying Feeling Rules

The central analytical distinction in this article is between an emotion that many people happen to feel and an expectation about what should be felt, shown or withheld. We treated four kinds of evidence as indicative of the latter: respondents formulating an expectation in normative terms rather than merely reporting a feeling; an identifiable source that communicates or reinforces it; a perceived consequence for departing from it; and a discernible gap between what respondents said they felt and what they said could be shown. None of these is a necessary condition, since feeling rules can operate without any nameable enforcer, and their absence is not decisive. Taken together, however, they allow the four clusters to be compared rather than asserted, and that comparison is reported in Section 4: the four are not equally well supported as rules, and we say where each is strong and where it is thin.

3.7. Relationship to a Prior Publication from the Same Study

This article and Cocîiu and Rughiniș (2025) draw on a single programme of doctoral fieldwork in the same commune, and the relationship between them should be stated precisely. The earlier article drew on long-term participant observation, fieldnotes and photographic material, together with the five semi-structured interviews conducted in June and July 2025. The present article draws on twenty interviews, of which those same five form a part; the remaining fifteen were conducted some ten months later, in April and May 2026, and were not analysed or reported in the earlier article, and the observational and autoethnographic material that formed the core of the earlier article is not used here. No interview quotation appears in both articles. The two analyses differ in question, framework and findings: the earlier article asked how a community makes sense of drought and infrastructural failure, using Gusti’s frames and manifestations together with Norgaard’s organised denial and Nixon’s slow violence, and reported themes of crisis normalisation, climate silence, institutional distrust and structural inequality. The present article asks how the emotional demands of scarcity are organised and distributed within households, using Hochschild’s framework, and reports feeling rules and their allocation. The present article is therefore a partially secondary analysis of a shared corpus, extended by fifteen new interviews, and not a re-presentation of published findings.

4. Results

4.1. Shame and Face-Work in Encounters with Outsiders

The first cluster concerns shame. When households without water receive visitors from places where water flows normally, the host must manage the gap between what domestic hospitality ordinarily requires, including a working toilet, water for handwashing, and the ability to offer a glass of water, and what the household can provide. Stated normatively, the expectation is that one should feel embarrassed before outsiders and should work to minimise the visible signs of water poverty. This was a recurrent pattern, though respondents used different words for the feeling and different strategies for managing it.

A private-sector employee (Int. 6) located it in the contrast between two worlds:

“Yes, yes, yes. Especially when I receive a guest who comes from other parts of the country, or from cities where water flows without limit. Here, I feel absolutely embarrassed [mă simt absolut jenat]. I have to say that you must be careful, that is, not to let the water flow without limit, to be careful to manage every liter of water.”

(Int. 6)

Shame here arises not from the absence of water as such but from the encounter between two material conditions, made visible by a person who embodies the norm from which the household departs. In Goffman’s (1959) terms, the guest threatens the host’s face by exposing the distance between the identity the household claims and its actual conditions. The work is double, since feeling and situation must both be managed, and the instruction to manage every litre is at once practical necessity and emotional discipline.

A healthcare worker who had recently moved into a newly built house (Int. 10) attached the same feeling to the house and to the body:

“When we moved into our new house without water, and family and friends came to visit, it was shameful [este rușinos] not to be able to flush the toilet or wash our hands. I mean, these are basic things…”

(Int. 10)

A new house carries a promise of comfort and mobility, and when it lacks running water the distance between promise and reality becomes a source of shame activated whenever visitors arrive. The bodily references locate the feeling at the level of minimal conditions of dignity rather than optional courtesy. Asked directly whether she had felt embarrassed, she answered affirmatively, and she said elsewhere that had she known the conditions the household would not have moved.

Her husband (Int. 9), who discovered the problem only after arriving, described the same encounter from the newcomer’s position:

“Yes, that is what they would say: ‘But you bought land here, and there is no…’ Well, yes, but I told them we were close to the town, and they said: ‘How is it possible that there is no water here?’”

(Int. 9)

The visitors’ question functions less as inquiry than as judgement, marking the commune as failing a basic expectation of settled life. He considered the situation “rather shameful”, particularly given its proximity to a town, which makes the absence harder to attribute to isolation and therefore harder to normalise.

The clearest evidence that this is an expectation and not only a feeling comes from a retired man (Int. 8). Asked whether he had felt embarrassed, he rejected the interviewer’s term and substituted a stronger one:

“Not embarrassed, but ashamed [Nu stânjenit, dar și rușinat?]. Well, the situation, you know…”

(Int. 8)

The correction is directed at the interviewer’s wording rather than his own. Feeling rules specify not only which emotion is appropriate but its intensity (Hochschild 1983), and here a respondent declines a milder characterisation offered to him and insists the situation warrants the stronger term. Asked separately how people in surrounding localities react on hearing about the commune’s water situation, he identified both a source and a penalty:

“It offends us. They call us stupid [Ne fac proști], both me and all of us.”

(Int. 8)

This is where the cluster is most clearly a socially organised expectation: there is an identifiable source, residents of neighbouring communes, and a specifiable consequence, being characterised as collectively incompetent. “Stupid” indicates that outsiders read the absence of water as a failure of the community rather than of institutions, and the shift from “me” to “all of us” shows the judgement absorbed collectively.

A young man (Int. 1) registered the feeling bodily rather than evaluatively: “Of course, also nervous tension, a hollow in the stomach [un gol în stomac]. I do not even know how to explain these feelings.” The inability to explain marks a limit of the emotion work, where the feeling exceeds what verbal accounting can manage. He also extended the expectation beyond the guest encounter: “It is also shameful [rușinos] in this day and age to go to the river. Nobody goes anymore. Twenty years ago perhaps they still went, but given the year we are in, to go to the river? It is shameful.” Collecting water from the river, once ordinary, is now coded as backwardness, and the coding is explicitly historical rather than timeless.

Others used a milder register. A young woman caring for small children at home (Int. 4) said: “Every time we have guests I feel bad [mă simt prost], so to speak.” The Romanian adjective carries connotations of both “bad” and “stupid”, and “so to speak” signals her awareness that it does not quite fit; this is a colloquialism of shame in Chase and Walker’s (2013) sense. The distribution of vocabulary in our corpus supports their observation, with a caution: the strong stem rușin- appears in five of the twenty interviews and the milder register of jenă and stânjenit in fourteen, but the milder term was frequently supplied by the interviewer’s own question, so the asymmetry partly reflects the instrument. What is more telling is that respondents rarely escalated unprompted, and that the one who did (Int. 8) did so by rejecting the milder term.

Two respondents departed from the pattern in instructive ways. A woman in her forties working at home (Int. 11) reframed the encounter as thwarted preparation rather than shame: “It was not so much [shame], generally, as the inconvenience: perhaps certain things needed doing and I could not do them in time or prepare as I would have liked because of the lack of water.” Her account links external judgement to domestic competence, but the transcript does not establish that this reframing conceals shame. She was also the only respondent to speak of discrimination, and she used the term for inequality inside the village rather than for the commune’s standing elsewhere. An education-sector employee (Int. 3) denied the experience outright: “No, no, there have been no such situations.” The transcript cannot distinguish an absence of shame from reluctance to discuss it, so we treat the response as a negative case rather than as evidence of concealed feeling. Asked about the future, she said: “Honestly, I see it as it is now. Absolutely not at all.” Together, these brief answers document departure from the dominant pattern without allowing an inference about her inner emotional state.

The shame cluster is strongly supported by three of the four indicators set out in Section 3.6: respondents formulate it normatively, an external source is identifiable, and departure carries a stated consequence. Evidence of a private/public gap is thinner, because neither reframing nor denial can establish what a respondent privately felt. The cluster is also gendered in ways Section 4.4 examines.

4.2. Worry, Helplessness, and the Emotional Texture of Running Low

The second cluster concerns worry. As reserves diminish, respondents describe stress, anger, nervous tension and a bodily sense of powerlessness. Unlike shame, which an outsider’s presence triggers, worry operates within the household, structuring relations between members and producing recurrent episodes tied to the rhythm of water availability. The expectation at issue is not primarily about what should be felt but about who must absorb the anxiety of scarcity so that daily functioning can continue, discharged through monitoring reserves, rationing, managing others’ reactions, and containing one’s own distress.

A young man who had returned from abroad (Int. 1) described worry becoming household conflict:

“Of course, a great deal of nervousness, because, well, it is abnormal not to have access to water. Discussions, arguments, because, well, you express your point of view—sometimes to your husband, sometimes to your wife, sometimes to your brother—because, well, nervousness and dissatisfaction inevitably arise.”

(Int. 1)

“Nervousness” names worry as a bodily condition rather than an appraisal; “inevitably” marks the quarrels as structural rather than as a failure of family harmony; and the list of family members shows the discharge distributed across household relationships. Wutich (2009) documented a similar pattern in urban Bolivia, where distress was expressed within the household through anger and interpersonal tension rather than collective action.

An education-sector employee (Int. 2) described its temporality:

“Yes, stressed most of the time; it may happen that [the truck] does not arrive that day either. We have to wait until the next day, when the truck arrives, and sometimes, as you can imagine, because we are a large family, we have to use water sparingly.”

(Int. 2)

The stress is continuous rather than episodic and organised around waiting for a delivery that may not arrive. Using water “sparingly” marks the discipline: consumption governed by calculation rather than need, which is itself emotion work because it requires suppressing the impulse to use water freely. He located the weight of running low on his mother, describing her as the first to become upset when water runs out while laundry and bathing still have to be managed. Her anger attaches to the impossibility of completing tasks that are her responsibility, consistent with Hochschild’s (1989) account of the second shift. He also placed the household in a wider geography of deprivation, saying they wait for the truck “as if we were in Africa”, where “people there rejoice when they see water. We are the same”—a comparison that simultaneously normalises the situation and registers its abnormality in a European context.

Powerlessness recurs as a key term, and one respondent gave it two objects: “We can feel powerless before God, asking Him to help us, and also before the authorities, who do absolutely nothing. And I know that life is vital; it revolves around water” (Int. 6). The doubling locates the household between two orders that should provide relief and do not, the theological and the institutional.

Worry also required justification. The respondent working at home (Int. 11) said: “It bothers me, first of all, because it also affects the quality of life. I cannot carry out all the activities I would like to do… Water is a necessary thing, it is not a whim [nu este un moft].” The final clause pre-empts an imagined accusation of excess, and producing it is itself emotion work, since she must manage not only her frustration but how it will be received, including by the interviewer; having to demonstrate that one’s needs are reasonable rather than extravagant is a feature of what Sayer (2005) called the moral economy of class. A retired woman (Int. 7) used a simpler vocabulary, “I get upset, I get upset [Mă supăr, mă supăr]”, and, asked whether her anger was directed at the situation or the authorities, redirected it toward collective human action: “this drought is probably due to us, we cut forests… I think the greatest blame belongs to humans.” Where most respondents located responsibility institutionally, she located it in shared failure, yielding guilt rather than helplessness.

The uneven arrival of the network gave worry a spatial dimension. A service-sector employee (Int. 12) said:

“Obviously, it is an ugly situation. It is not normal for us not to have… Especially since we are on a side lane and, effectively, one part of the village has water; one part of the village does not have water. And it is not okay. That is, everyone pays taxes, we pay the same, regardless of whether we are on a side lane or on the main road.”

(Int. 12)

The contrast between lane and main road maps access onto a spatial hierarchy, and the appeal to equal taxation frames the response as justice rather than inconvenience, corresponding to Anand’s (2017) observation that connection distributes recognition as well as water. The same respondent described designing household infrastructure around anticipated shortage, installing a shower rather than a bathtub, so that the physical environment enforces the discipline of conservation.

Worry differs from shame in its temporal structure, being chronic rather than episodic, and it differs in something more consequential for our argument. Assessed against the criteria in Section 3.6, worry is the least rule-like of the four clusters. Respondents describe it vividly and consistently, but they rarely formulate it as an expectation about what one ought to feel; no respondent identified a source communicating such an expectation; and none described a consequence for failing to worry. Worry as a prescribed feeling is thinly supported, and we do not claim it. What the accounts do support is an expectation about its allocation: that someone in the household must absorb and contain it, and that this someone is not randomly selected. That expectation is stated normatively, it is reinforced by the domestic division of labour, and it is visible in the accounts above, where the person who discovers the shortage and reorganises around it is consistently the person responsible for the tasks requiring water. Since Hochschild’s rules govern the management and display of feeling and not only its occurrence, an expectation of this kind falls within the framework; but it concerns who does the emotional work rather than what anyone should feel, and the distinction is worth marking rather than eliding.

4.3. Hope as Social Obligation and Resignation as Contained Expression

The third and fourth clusters operate as a pair. The hope expectation prescribes maintaining and publicly expressing the belief that the problem will eventually be solved; the resignation expectation permits expressions of defeat, but only in contained, formulaic terms that do not escalate into open anger or protest. Together they manage the temporal horizon of water poverty: hope keeps the future open, while resignation prevents repeated disappointment from becoming socially disruptive.

The proverb “speranța moare ultima” [hope dies last] appears in two of the twenty interviews. It is not, therefore, a widely circulating formula in our material, and we do not claim that it is; what makes the two instances valuable is how they are used. A young woman with small children (Int. 4) offered it as her entire answer to whether she still hoped the situation would improve:

“Hope dies last.”

(Int. 4)

Three observable features matter: the response is complete in four words; it is neither elaborated nor qualified, and no grounds are offered; and the interviewer, receiving no opening, moved immediately to an unrelated question about who fetches water. The formula closed the topic. Whether it also expressed hope is not something the transcript can settle, and we make no claim about it.

The second instance is more explicit. A service-sector employee (Int. 12) marked the proverb as a citation before undercutting it:

“You know, as the saying goes [cum e vorba aia], hope dies last, but even that eventually dies… I would have been glad to have water by now. We used to have rain; now… nothing. Let us see some movement, more movement, not just this much.”

(Int. 12)

“As the saying goes” is a metadiscursive marker: the speaker presents what follows as a conventional expression and immediately qualifies it. This is direct evidence, requiring no inference about inner states, that a speaker can explicitly cite a shared formula while withholding full endorsement. The remark about rain also marks a loss institutional action cannot repair, so that the ecological dimension exceeds the institutional remedy.

At the opposite pole sits a formula of self-blame. The retired man (Int. 8) used “ne merităm soarta” [we deserve our fate] six times in varying forms. Asked whether the situation was unjust, he first identified institutional failure, saying the person placed in office had attended to something else, then immediately added: “We deserve our fate. As long as we are how we are and if we will continue to be like this, we deserve our fate.” Elsewhere the reasoning is explicit, and it is electoral rather than fatalistic: “who puts the officials in the town hall? Who votes for them? And those at the county, and those in Parliament, who? Who? Not us? And do we deserve our fate?” The formula is not passive acceptance: it contains anger by distributing blame between the authorities and the community that elected them, so that it finds no single target, and it names the situation in terms that close the conversation rather than opening it toward protest. Stambe and Parsell (2024) described a comparable channelling of feelings that might motivate collective action into self-directed moral evaluation. He also rehearsed the formula privately, “And I say to myself: I deserve my fate, I deserve my fate”, so it operates in internal management and not only in interaction.

Others described hope’s decline without formulae, and their phrasing is instructive. The respondent working at home (Int. 11) said: “At the moment, given the situation that has been going on for some years now, I would like to believe it will be resolved, but I have rather lost hope.” “I would like to believe” states a desire to believe rather than a belief, and “rather” softens the admission; even in declaring hope’s failure she performs partial compliance. Admitting failure outright appears easier once conditions have changed. A retired woman whose household was connected in the autumn before her interview (Int. 15) said of the preceding period: “It is true, yes, yes, that is how it is. I no longer believed such a thing was possible.” The past tense marks a period in which the expectation had been abandoned, and she speaks from partial resolution; she also specified how limited the change was, “there are very few houses, from the entrance of the village up to the town hall, that are connected. That is, a quarter of the village.” The teacher who denied shame (Int. 3) offered the starkest break, saying of the village’s future: “Honestly, I see it as it is now. Absolutely not at all.” She does not hedge, soften, or invoke the proverb. The brevity documents a refusal of optimistic framing in this interview, but it cannot establish whether she had withdrawn from a wider social expectation or was making only an assessment of infrastructure policy.

Resignation has its own formula, and its use is revealing. Asked what she would tell someone from outside the commune about life there, a healthcare employee (Int. 13) answered: “We are content with things as they are; we settle for it [ne complăcem].” The interviewer echoed the phrase and she completed it: “In the situation we are in, and we make our own sources.” Two features matter. The formula was produced unprompted and specifically as an account for outsiders, which links resignation to the shame expectation of Section 4.1, since what one says to outsiders about the situation is itself governed. And the resignation is not passive: “we make our own sources” indicates practical adaptation, in which the community has given up not on having water but on obtaining it institutionally, redirecting effort toward wells, purchases and rainwater collection. Collins et al. (2024) describe such strategies as substitutes for institutional supply that become normalised and reduce pressure on authorities to act.

A private-sector employee (Int. 17) observed the pattern from outside it, noting that “some people have resigned themselves” and that the priest “connects it to God and to sins”, so that drought is framed by religious authority as divine punishment. This supplies a further source: if the situation is God’s will, resignation becomes a morally appropriate posture rather than merely a pragmatic one. A retired woman (Int. 7), asked whether she still had hope for the water project, relocated it rather than abandoning it: “Minimal, very small. We are old, but we would like it for these children who are here.” Transferring hope to a future she will not share preserves the expectation while acknowledging its personal exhaustion.

Assessed against the criteria in Section 3.6, these two clusters are supported differently from worry. For hope, respondents formulate the expectation normatively, describe a cost to declaring defeat, and display a gap between private and public positions, most clearly in the hedging of Int. 11 and the retrospective admission of Int. 15, though the source is diffuse, residing in ordinary talk rather than in an identifiable enforcer. For resignation the constraint is on form rather than content, and religious framing supplies an additional source. Neither rests on a claim about what respondents inwardly felt.

4.4. The Gendered Distribution of Emotion Work

These expectations do not operate equally across household members. The interviews indicate a recurrent pattern in which women carry a disproportionate share of both the practical labour of water management and the emotion work of living under scarcity, following the logic Hochschild (1989) described. The basis for that statement should be given precisely, because the denominator is not twenty.

The question whether water scarcity affects women and men differently was put in fifteen of the twenty interviews, not in the five earlier interviews described in Section 3.4, which used a guide developed for a different research question. Of the fifteen asked, thirteen affirmed that it does. The two who did not were both men, and neither denied a difference: one said “I have not thought about that” (Int. 9) and the other “I do not know, I could not answer that” (Int. 12). Of the eight women asked, all eight affirmed a difference; of the seven men, five did. The grounds given cluster on responsibility for domestic tasks and on bodily hygiene.

The most developed account came from an education-sector employee (Int. 2):

“It depends greatly here. Most likely, yes, because we have different habits. Women mostly stay at home and need a very large amount of water for all the activities they carry out. I mean women who stay at home, have children, and take care of the household… while men leave for work, come home in the evening, use the water, take a bath or shower, and continue their activities the next day. Meanwhile, women stay at home and need water.”

(Int. 2)

Its value is that it makes the temporal structure explicit. Men’s relation to water is compressed into the evening; women’s spans the day, because the tasks for which they are held responsible require water continuously. Women therefore encounter the conditions described in Section 4.2 with greater frequency and duration: they monitor reserves, discover that water has run out, and reorganise accordingly. The same respondent described his mother as the first to become upset when water runs out while laundry remains, locating the expression of worry on a specific person, not as an individual trait but as a consequence of her position in the division of labour.

A retired man (Int. 14) described the division as self-evident: “We, when we come home, throw the clothes and the wife has to wash them for us. So, everything starts from the hygiene issue, everything is about hygiene.” “Throw” casts men’s relation to dirty laundry as disposal, and “has to” marks washing as obligation rather than choice; asked who carries more water in the community he answered without hesitation, “Women more.” He then volunteered a qualification that complicates the picture, and it is the most analytically interesting statement in this part of the corpus:

“What happens is that in the countryside the wives are mostly housewives, but, attention, because now all the women have begun to be employed as well. So both come home in the evening and now that thing disappears… the housewife.”

(Int. 14)

Here a respondent historicises the division he has just described as obvious. The arrangement in which women are continuously available to domestic water tasks is, on his own account, being eroded by women’s paid employment. The consensus that women are more affected is therefore not a description of an unchanging order but of an arrangement respondents themselves know to be in transition, and one that may intensify rather than relieve the burden where employment is added to unchanged domestic responsibility.

Another retired man (Int. 8) stated the allocation compactly: “Consumption… the wife takes care of it, the supply has fallen on me.” The division between consumption, meaning management of use inside the household, and supply, meaning procurement from outside, maps onto an interior and exterior distinction, and his verbs are worth noting: consumption is something his wife “takes care of”, implying continuous management, while supply “has fallen on” him, implying an obligation that arrived rather than one chosen. He also illustrates why the denominators matter. Asked whether scarcity affects women and men differently, he first equalised, “Well, it affects both me and her”, and only after a follow-up asking who is affected more did he answer: “Her, because she is the first to jump, to shout that there is no water… loosely speaking, shouting. That we have no more water. I go quickly.” The gendered claim emerged on probing rather than spontaneously, and what he described is precisely the position of first encounter.

Where men described the pattern they sometimes did so with more assurance than the women concerned. The private-sector employee (Int. 6) said: “I see my mother as very affected, especially during the summer, because she does not have enough water for the animals, for the garden, for washing. I know very well that women keep the household, in the end.” A retired woman recently connected to the network (Int. 15) hedged the same claim: “Yes, perhaps we want more water, well, in the household, when you cook, when you do other things.” Her “perhaps”, and her choice of “want” rather than “need”, sit oddly against the confidence with which male respondents asserted women’s greater requirement, and the contrast is worth registering rather than smoothing.

This bears on how the hygiene material should be handled. Several respondents explained women’s greater need by reference to the body. A participant (Int. 5) said: “Women, well, given that they have to wash their hair… depending on how long the hair is, well, if we have a habit of washing hair twice a week, let us say, we try to wash it only once, and if we don’t have water, well, we adapt.” Another put it more briskly: “Women, yes, have long hair. They also have to bathe. Automatically, more water is consumed” (Int. 4). These are evidence of locally held norms about gender and bodily maintenance, and of the discipline those norms impose under scarcity; they are not evidence that women require more water physiologically, and we do not treat them as such. A socially specific standard of feminine presentation becomes a site at which scarcity is absorbed, and the absorption is described as adaptation rather than loss. The accumulation of such adjustments reduces bodily comfort disproportionately for women, but the mechanism is normative rather than natural. Similar accounts recurred across the sample: a healthcare employee described the responsibility as “washing laundry, cleaning first of all. Keeping the household matters very much” (Int. 13); a retired woman began unqualified and then weakened it, “the woman has to cook, the woman has to wash… although the husband also puts the laundry in the machine and can also…”, the qualification trailing off (Int. 7); and a young man reproduced the same account as the retired respondents (Int. 1), which suggests the perception is not a generational artefact.

The one household represented by two respondents permits a limited comparison, instructive on a specific point rather than a general one. The wife (Int. 10) described the burden in detail including its emotional dimension, saying of the household atmosphere “Yes, it spoils the mood”, and describing withdrawal to her parents’ home when conditions became too difficult, a spatial form of emotional management. Her husband (Int. 9) said he had not thought about whether scarcity affects women and men differently. Within a single household, then, the same conditions are described as gendered by one member and as unconsidered by the other. This does not demonstrate that emotion work is unequally distributed there, and we do not offer it as proof of the distribution; it illustrates that the distribution is differentially visible depending on one’s position within it, consistent with Wutich’s (2009) finding that men were less likely to recognise the disparity.

Two qualifications follow. The evidence for women’s disproportionate share of practical water-related labour is strong and consistent, attributed to women across age groups, occupations and both genders. The evidence that women additionally manage other members’ feelings is thinner, resting on a mother’s anger when tasks cannot be completed (Int. 2), the allocation of continuous consumption management to a wife (Int. 8), one wife’s account of household mood and her withdrawal from it (Int. 10), and one description of a woman as first to react to shortage (Int. 8). These point consistently in one direction but do not establish that women systematically manage others’ emotions on the household’s behalf, and we do not claim it. These are also for the most part reports about how households work rather than observations of them, offered by people occupying different positions within the arrangements they describe.

What the accounts do support is that women are positioned to encounter shortage first and to reorganise around it; that the continuous work of rationing is attributed to them and the episodic work of procurement to men; that the bodily costs fall on them through locally specific norms of presentation; and that at least some emotional consequences of inadequate supply are described as theirs to absorb. Reay’s (2000) notion of emotional capital, the affective resources invested in maintaining family functioning without being recognised or replenished, describes the resulting position, and under chronic scarcity the depletion is continuous rather than episodic.

5. Discussion

The analysis identified four clusters of expectation governing emotional conduct: shame activated by outsider encounters, the absorption of worry activated by falling reserves, hope requiring public maintenance of optimism, and resignation permitting defeat only in contained terms. These are not individual coping styles but socially organised patterns, widely represented across the interviews despite variation in age, occupation and duration of residence, and they are not equally well evidenced. This section develops what follows from them.

5.1. Infrastructural Emotion Work

In Hochschild’s (1983) framework the demand for emotion work comes from an institution: the airline requires warmth from flight attendants, the bill collector must produce coldness, the family expects the wife to maintain the emotional climate of the home. The source is always social in a specifiable sense, whether an employer, a market, or a set of gendered expectations embedded in family structure. Our findings indicate a further source. Emotion work can also be generated by infrastructure, or more precisely by the sustained failure of infrastructure to provide what it is socially expected to provide. We propose the term infrastructural emotion work for the unpaid emotion work this produces.

The point requires care, because it would be easy to overstate. It is not that matter issues instructions. No employer requires residents of this commune to feel ashamed when guests visit or to maintain hope after years of inaction, and in that limited sense these expectations are not imposed by any single institution. But they are thoroughly socially organised all the same, and by identifiable things: hospitality conventions specify what a host owes a guest; norms of domestic competence specify what a well-kept household looks like; expectations of modern infrastructure make the absence of a network anomalous rather than natural, which is why proximity to a functioning town intensifies the feeling rather than relieving it; gender relations allocate the resulting work; religious interpretation, in the priest’s framing of drought as punishment for sin, supplies a further warrant for resignation; and the judgement of outsiders in neighbouring communes supplies the sanction directly, as Section 4.1 showed. Infrastructural failure does not create these expectations from nothing. It activates them, gives them a daily occasion, and makes compliance costly.

Understood this way, the concept extends Hochschild’s framework toward the study of infrastructure while remaining distinct from existing work at that intersection. Star (1999) established that infrastructure becomes visible in breakdown; we specify one thing that becomes visible, namely the distribution of emotion work that keeps breakdown liveable. Rodgers and O’Neill (2012) showed how marginalisation operates through material arrangements; infrastructural emotion work names a cost of that operation borne privately and left unrecorded. Bosworth (2023) distinguishes two established senses of affective infrastructure, infrastructure that generates emotion and affect that organises politics; ours is a third question, concerning the work of managing feeling and the expectations governing it. Carr et al. (2026) describe an infrastructural politics of emotion in which people anticipate discomfort and develop routines within community infrastructure they inhabit; our setting differs in that the expected infrastructure is missing or unreliable, so routines develop around a gap. The concept has purchase beyond water: wherever heating, sanitation, electricity or transport fails chronically, households must manage the distance between expected and actual conditions, and that work is uncounted, uncompensated and unrecognised in exactly the way Hochschild showed domestic emotion work to be.

5.1.1. Socio-Material Generation of Feeling Rules

In Hochschild’s (1979, 1983) formulation, feeling rules are cultural prescriptions learned through socialisation and enforced interactionally: one should feel happy at a wedding and sad at a funeral. Our findings suggest a variant in which the same kind of expectation is generated socio-materially, at the meeting point of a material condition and a social standard, and it is worth being precise about the mechanism because the temptation to shortcut it is strong.

The absence of water does not by itself produce shame. It produces shame because hygiene, hospitality, domestic competence and modern living are socially valued, and because a household without water cannot demonstrate them before a guest who can. Remove the guest and the material condition is unchanged while the shame subsides; remove the hospitality norm and the same empty tap would carry no such charge. Similarly, the descent of water in a cistern generates worry partly because household members understand themselves as responsible for cleanliness, care and family functioning, and would be judged if those failed. What the material condition contributes is not the content of the expectation but its occasion, its frequency, and the cost of meeting it. This is why we describe these expectations as socio-materially generated, or materially activated, rather than as materially originated: matter does not establish normative expectations independently of social interpretation.

The uneven arrival of the network functions as a partial test, and the result is more interesting than a simple confirmation. If these expectations are tied to material conditions, altering the conditions should alter them, and to a degree it does: the respondent connected in the autumn before her interview could state retrospectively that she had stopped believing connection was possible (Int. 15). But connection did not dissolve the expectations. Households on the network reported that the water smelled strongly of chlorine and continued to buy bottled water for drinking (Int. 9, Int. 10), so a problem of quantity was replaced by one of trust in quality. And connection redistributed rather than removed the emotional stakes, since those on unconnected lanes now compare themselves not only to a town where water flows but to neighbours on the main road, a comparison one respondent called discrimination within the village (Int. 11). Anand’s (2017) account of connection as incremental, intermittent and reversible describes this better than a binary of access and its absence. The conclusion is accordingly weaker but more defensible: infrastructure reorganises the emotional conditions of domestic life rather than dissolving them, and a network arriving in part of a settlement can generate new comparisons even as it relieves old burdens.

5.1.2. Temporalities of Infrastructural Emotion Work

Domestic emotion work is often treated as a continuous, undifferentiated flow (Hochschild 1989). Our material suggests that under chronic scarcity it operates across distinct temporalities, each associated with a different expectation. The shame expectation generates episodic work, triggered by a guest’s arrival and involving preparation, management during the visit, and incomplete recovery afterwards. The worry expectation generates chronic work, tied to the daily rhythm of reserves and demanding ongoing vigilance and rationing. The hope expectation generates sustained public work, maintained over years regardless of whether conditions justify it. The resignation expectation generates work of a fourth kind, brief and closing, visible where a conventional response ends rather than extends an account.

These temporalities coexist in the same household and often in the same person, so that the total burden is not a single continuous demand but the sum of overlapping demands with different rhythms, triggers and costs. This connects to Nixon’s (2011) account of slow violence operating across multiple timescales at once, and to Gill et al. (2026), who found that resource loss in flood-affected communities produced psychosocial stress across different temporal horizons. Our addition is that these timescales are not merely experienced but actively managed, and managed differently.

5.1.3. The Configuration of the Four Expectations

The four expectations are related, and it is worth saying in what sense, since a stronger claim would not be warranted. We do not claim a causal sequence, and our data could not establish one. What the material supports is a configuration: the four co-occur, they operate in different situations and on different timescales, and they constrain one another’s expression. Shame presupposes monitoring, since one cannot be ashamed of conditions one does not track; the hope expectation constrains resignation, in that defeat may be conceded privately or in formula but not declared outright; and resignation in turn limits how far frustration can escalate toward open political anger. The resignation formula addressed to outsiders (Int. 13) shows the connection working directly, since what may be said about the situation to someone from outside is governed by the same concern with external judgement that produces shame. This is an analytical configuration rather than a model, and it is best read as describing how a set of expectations divides the emotion work of a situation between different occasions.

5.2. Formulaic Emotional Display: A Mechanism of Infrastructural Emotion Work

Hochschild (1983) distinguished surface acting, producing the outward signs of a feeling without experiencing it, from deep acting, working on the feeling itself. Our material draws attention to a conventionalised discursive form within this territory, which we call formulaic emotional display: a culturally available phrase that provides an interactionally acceptable response to a feeling rule while foreclosing further elaboration.

The claim is deliberately restricted to what interview data can support, and it is worth stating what we do not assert. We do not claim that a speaker using a formula feels nothing, or is not attempting to generate the feeling, or expends less emotional effort than someone who elaborates. None of that is observable in an interview, and an earlier version of this argument asserted it; the second author’s objection that the claim attributed an inner process to which we had no access was, on reflection, correct, and the reformulation here is the result. What is observable is a pattern of four features, each present in our material. The response is brief and offers no grounds, as when the proverb about hope constituted an entire answer (Int. 4). It closes the topic, so that the interviewer received no opening and moved on. It can be explicitly marked by the speaker as a convention, as in “as the saying goes, hope dies last, but even that eventually dies” (Int. 12), where the formula is immediately qualified. And it can be repeated across contexts and rehearsed privately, as the formula of deserved fate was, six times and in internal self-address (Int. 8).

Read this way, formulaic display is not a third mechanism alongside surface and deep acting but a conventionalised discursive form within the territory Hochschild mapped, available where hardship lasts long enough for shared phrasings to accumulate. Its interest lies in what it accomplishes interactionally: it provides a culturally recognisable response to an unresolved situation without requiring an elaborated personal account, and it can close off follow-up questions. That the formula addressed to outsiders concerns precisely what the community should say about itself (Int. 13) suggests the mechanism is oriented to external judgement, which links it to the shame expectation rather than to hope alone.

Two qualifications follow. The first concerns evidence: the hope proverb appears in two of our twenty interviews and the formula of deserved fate in one, so we offer formulaic emotional display as a mechanism identified in this corpus rather than as a widely distributed practice demonstrated by it, and establishing its prevalence would require a design suited to that question. The second concerns terminology. The phrase formulaic emotion work has been used before, by Hebson et al. (2007), for the prescriptive and standardised emotional conduct demanded of English teachers under capability procedures. Their construct is related but not identical: theirs is prescribed by an employer operating a performance regime, ours is a community repertoire with no institutional author. We adopt formulaic emotional display both to mark that difference and to keep the claim at the level of observable display. The mechanism also connects to Chase and Walker’s (2013) observation that shame circulates through colloquialisms rather than being named directly, and to Sayer’s (2005) account of class-related feelings managed through conventional moral vocabularies; in each case a shared vocabulary provides culturally recognisable forms through which speakers position themselves. It is in this sense that a community facing chronic infrastructural failure can be said to build an emotional infrastructure alongside the material one it lacks.

5.3. Gender and the Distribution of Emotion Work

Our findings support Hochschild’s (1989) account of the gendered distribution of domestic labour and add a specification, with a limit that should be stated plainly. The specification is that women’s position in the household division of labour places them at the point of first encounter with shortage: held responsible for cooking, laundry, cleaning and childcare, they discover that water has run out, reorganise tasks, and absorb the immediate consequences, and the continuous work of rationing is attributed to them while the episodic work of procurement falls to men (Int. 8).

The limit is that our evidence for practical labour is considerably stronger than our evidence for the management of other members’ feelings, and we have not claimed more than the material supports. Two further points guard against overreading. Respondents’ explanations of women’s greater need, particularly those invoking hair-washing and hygiene, are locally held norms about feminine presentation rather than physiological facts, and are analysed here as norms determining where scarcity is absorbed. And one respondent volunteered that the arrangement is changing as women enter paid employment (Int. 14), which suggests that what we document is a specific and possibly transitional configuration rather than a stable order, and that adding employment to unchanged domestic responsibility may intensify rather than relieve the burden.

5.4. Limitations and Directions for Future Research

Several limitations should be acknowledged. The study draws on a single commune, and the expectations identified here may not hold where cultural norms, household structures or relationships to water differ. Twenty interviews are adequate for identifying shared expectations but not for establishing how prevalent or intense any emotional response is, and no such claim is made. The design is not longitudinal and cannot trace how these expectations formed or change. It is worth being clear about what the two periods of interviewing do and do not allow in this respect. They straddle the arrival of the network, and the five interviews of 2025 were conducted as the first connections were being made, while the fifteen of 2026 followed some ten months later. But the two groups are different people, answering partly different questions, so the comparison is between two cross-sections of one community rather than a panel, and we have not treated it as evidence of change over time.

Three limitations concern the instrument and the analyst rather than the sample. The question of gendered impact was put in fifteen of the twenty interviews and not in the five that used the earlier guide, so that analysis rests on a smaller base than the study as a whole. The interviewer’s questions about embarrassment frequently supplied the milder Romanian term, so the distribution of shame vocabulary partly reflects the wording of the questions, and we have relied on unprompted escalations where the point mattered. And the first author’s membership of the community, which enabled the access on which the study depends, also made certain readings available too easily, particularly readings of formulaic speech as transparently meaningful; Section 3.3 describes how this was handled and Section 5.2 records where it changed a substantive claim.

Two findings we regard as open. The first is the consequence of connection: our material suggests that a network relieves the problem of quantity while introducing concerns about quality and generating new comparisons among neighbours, but four respondents in three households cannot settle what happens to feeling rules when infrastructure arrives. Longitudinal research following households before and after connection would address this directly, and the uneven commissioning documented here offers an unusually good natural setting for it. The second is the prevalence of formulaic emotional display, which our design identifies but cannot measure. Beyond these, comparative studies across communities with different durations of scarcity and institutional responses would test whether the configuration described here is specific to this setting, and infrastructural emotion work could be applied to chronic failures of heating, sanitation, electricity or transport.

6. Conclusions

This article began from a simple observation: when water stops flowing, households must manage not only the practical consequences of its absence but the emotional ones. The sociology of emotions has long recognised that feelings are socially organised, governed by expectations, and maintained through labour. What it has examined less is the part material infrastructure plays in generating that labour. Analysing how households in a water-poor Romanian commune produce and distribute emotion work, we identified four expectations governing emotional conduct: shame in encounters with outsiders, the absorption of worry within the household, hope maintained in public, and resignation expressed in contained terms. They do not carry equal normative warrant, and we have said where each is well supported and where it is not; the expectation concerning worry is the least rule-like of the four, and holds as an expectation about who must absorb feeling rather than about what anyone should feel.

The concept we propose is infrastructural emotion work: the unpaid emotion work produced when a material system that is socially expected to function does not. Three arguments elaborate it. These expectations are socio-materially generated, arising where a material condition meets a social standard, rather than prescribed by culture alone or dictated by matter alone. The work they demand is temporally differentiated, episodic for shame, chronic for worry, sustained for hope and closing for resignation, with all four operating at once. And one of its characteristic mechanisms is formulaic emotional display, in which a conventional phrase provides an interactionally acceptable response while foreclosing elaboration, a claim we restrict to what is observable in talk. The resulting burden falls unevenly, and women’s position in the household division of labour places them at the point of first encounter with shortage.

One implication deserves emphasis in a form the evidence supports. Installing a water network is not only a technical intervention but an emotional one, because it alters the conditions that make shame, worry, hope and resignation daily concerns. But our material cautions against treating it as a solution to the emotional situation. In this commune, the network arrived along the main road and not the side lanes; connected households distrusted the water that came and kept buying bottled water; and the comparison between connected and unconnected neighbours made a new form of grievance visible in the interviews. Infrastructure reorganises the emotional conditions of domestic life rather than dissolving them.

The concepts developed here are not limited to water. Wherever material systems fail chronically, households must perform emotion work to manage the distance between expected and actual conditions. Infrastructural emotion work names that work and opens it to analysis; and where pipes do not deliver, households build an emotional infrastructure of their own.