We take it for granted that social research methods, both qualitative and quantitative, give us access to social lives. Interviews seem to let us hear people’s lived experiences. Surveys and datasets seem to show us how people in society live their lives. And when we combine both, we imagine we’re getting the best of both worlds: a richer, more balanced picture of the truth. But what if all of these methods are actually operating at the same level? What if, beneath their differences, they are all just different ways of working with traces: compressed and filtered symbolic representations of life that have already been shaped before we ever begin to study them?
This idea forces us to rethink the ontology of our data. It’s not just that research works with symbolic representations, it’s that those representations have gone through a long process of formatting, translation, and stabilization before they ever even become “data.” They are never “first-hand.” They never simply reflect life as it is. Instead, they stand in for life after it has been made fully legible, after it has been turned into something that can be asked about verbally, categorized, counted, or narrated.
To understand this, we need to think more about where life actually happens. Life unfolds not in theory, or in data, but in what I call the mesocosm: the everyday middle zone of how the world appears to us, where the world makes sense. It’s the space where bodies interact with environments, where people respond to each other, where places shape behavior, where tools and language guide action. Everything that matters, be it relationships, emotions, habits, rituals, struggles, joys, happens in this middle zone. Microscopic lives don’t make any sense to us unless they are rescaled into mesocosmic size. Macrocosmic forces don’t make any sense unless they are made graspable in multisymbolic form. And yet, this being-in-the-middleworld is what no research method could ever capture directly. Methods can only collect the symbolic traces left behind.
For something to become an object of social science inquiry, it must be formatted in ways that allow it to be stabilized and repeated. This process of formatting is everywhere: hospitals have to code diagnoses before we can talk about “hospitalization rates.” Schools have to standardize tests before we can talk about “learning outcomes.” Governments have to define categories like “employment” or “poverty” before they can be measured. And even in qualitative research, which seems more open-ended, there is always a reliance on these earlier forms of symbolic stabilization. When I study “depression in India,” for instance, I’m not just collecting interviews, I am also relying on medical categories developed by the WHO, on policy definitions set by ministries of health, on diagnostic protocols in clinics, on sales records by pharmacies, and so on. Even the most grounded ethnographic work draws heavily on preexisting categories that shape what kinds of answers can be given in the first place. I can’t do an ethnography of “depression in India” without “depression,” “India,” “psychiatry,” “pharmaceutical,” etc. having been fully stabilized in symbolic form.
Quantitative methods depend on this stabilization even more. For a survey to work, the phenomena being studied must already be shaped in ways that allow single questions to produce single answers. Satisfaction has to be something that can be measured on a scale from 1 to 5. Symptoms have to be clearly defined so they can be ticked off. These methods don’t reflect raw experience, they reflect life after it has been processed, narrowed, and formatted for legibility. In that sense, they are only possible when institutions, technologies, and prior knowledge systems have already done most of the work of shaping life into something countable.
Qualitative methods seem to offer more flexibility, and they do, at least up to a point. What they often capture is the early stage of that same symbolic stabilization process. They catch people in the act of trying to make sense of something, trying to put it into words, struggling with contradictions, pauses, hesitations. This is powerful. But even then, what we’re seeing is not unmediated life, it’s the beginning of symbolic stabilization. What the person says in the interview is already a translation of their experience into a form that fits the structure of the conversation, the expectations of the setting, and the language available to them.
In short, both qualitative and quantitative methods rely on symbolic rescalings of mesocosmic life. The difference lies in timing. Quantitative methods tend to work with fully stabilized categories: things that have already been formatted and fixed. Qualitative methods often work with categories still in motion: things that are just beginning to take shape. But neither method gives us direct access to the raw middle of life, that space of real-time coordination where meaning, value, and experience are actually generated.
This also means that all the long-standing debates between qualitative and quantitative research miss the point. The real divide isn’t about richness or rigor. It’s about how far along the symbolic stabilization process the data already is. Both approaches rely on what parts of life can be made legible,what can be named, described, measured, or narrated. And both struggle to deal with the parts of life that can’t be made legible in this way, those moments of silence, confusion, misalignment, and contradiction that resist easy translation.
Once we recognize this, we can also see why so many disciplines end up divided into false oppositions: structure versus agency, individual versus society, mind versus body, biology versus culture. These are not real splits in the world. They are the result of how we’ve carved up knowledge into different domains, each with its own methods and assumptions. The more a form of life resists legibility, the more likely it is to get misclassified, misunderstood, or simply ignored.
So what does this mean for the future of research?
First, we need to stop thinking that social research methods give us access to being-in-the-world. They don’t. What they give us is access to the symbolic traces left behind by life that has already been shaped into a form that fits our categories. This isn’t a failure, it’s a condition. But we need to be honest about it. Note that this is not another version the “everything is socially constructed” argument at all. We really live in the mesocosm and the mesocosm is the only really real there is. It’s not socially constructed. What I am saying is that any social science argument about the mesocosm can only draw on symbolic traces that living-valuing-remediating processes have left behind.
Second, we need to develop research practices that start with the question: what has already been made legible here, and what remains illegible? Which parts of life have been flattened to fit into categories, and which parts are still resisting that process? What infrastructures,hospitals, schools, bureaucracies, disciplines,have already shaped what can be asked, answered, and known?
Third, we need to pay more attention to how our methods don’t just study the world: they can change it. Every act of measurement, every interview, every data set reorganizes the field it studies. Research is not just describing the world; it’s actively reshaping the ongoing coordination of life. And the more we understand this, the more we can design methods that are responsive, respectful, and truly attuned to the richness of the middle world we all live in.
This is also not another argument about “response effects” or about a need to “be reflexive,” it goes much deeper. Consider the case of a researcher who studied a sacred healing ritual among an Indigenous community in Eastern India. The ritual is not “symbolic,” it IS healing, it IS life-renewing. The words and acts of the healer are not symbols and metaphors, they change the world. The researcher, clinging to Durkheimian theory, interpreted the ritual as a mechanism for social cohesion,reinforcing “group solidarity” through “collective effervescence.” But when the Indigenous people got to read the publication, they felt deep anger and grief. Elders said the analysis had cut open a world that was never meant to be dissected analytically. Words that carried living power (incantations invoked ancestors, spirits, and healing forces) were reduced to human-human “relatedness.” The ritual, for them, was not a performance "about" anything to do with "society," it was an act of worlding. By turning their living mesocosm into an abstract science experiment, and by fossilizing living-valuing into written symbols, the researcher violated the very mediational ground that the ritual sustained. In the community’s terms, the research had not interpreted the ritual, it had desecrated it. No amount of “reflexivity” by the researcher can prevent this, indeed ill-understood “reflexivity” exacerbates the harm. This was violent extraction, but it could not be mended by making people sign more elaborate informed consent forms. The violence of symbolic extraction cannot be addressed by further layers of symbolic extraction.
Ultimately, the most important question is not whether our methods are valid. It’s whether our writings help life become more livable, or, at least, whether we do not harm other people’s mesocosm through symbolization, inscription, and abstraction. That is the kind of reflexivity we should be aiming for. Not pushing for more and more legibility, but more care for the worlds that do not want to be described by us.