Progress in Human Geography
2022, Vol. 46(2) 652–671
© The Author(s) 2022
Article reuse guidelines:
Ambiguous spaces, empirical traces:
Accounting for ignorance when
researching around the illicit
, Karly Marie Miller
, Juliet Lu
, Lauren S Withey
University of Wisconsin-Platteville, Platteville, WI, USA;
Santa Barbara, CA, USA;
Cornell University, Ithaca, NY, USA;
San Francisco, CA, USA;
Astoria, OR, USA
Human geographers investigating socio-environmental change in resource frontiers often encounter illicit
activities occurring alongside the licit processes they study. These encounters pose logistical challenges to
conducting research and moral and analytical dilemmas for researchers. Illicit activities produce what we refer
to as spheres of ambiguity, which obscure certain information, relationships, and phenomena surrounding
them. We address the dearth of analytical tools for approaching the uncertainty this generates by bringing
concepts from agnotology—the study of ignorance—to bear on research conducted amidst illicit activity,
offering a framework for systematically evaluating ambiguous ﬁeld data that is incompletely observed or
illicit, ambiguity, political ecology, research methods, agnotology, knowledge, frontiers, ignorance
Cocaine smuggling. Opium, coca, and cannabis farms.
Drug money laundered through otherwise legitimate
plantations, ranches, and tourism enterprises. Research
informants who disappear for days or weeks with
vague explanations of working “over the hill”or
“up the river.”Warnings of when and where not to
travel. Encountering illicit activity during ﬁeld re-
search is remarkably common and oftentimes inevita-
ble. However, due to safety and ethical considerations,
methodological challenges, and a variety of other
obstacles, researchers working amidst illicit activities
often censor or avoid these topics both in the ﬁeld
and in writing (Hall, 2013). Navigating encounters
with illicit activities that are not the subject of re-
search can create notable absences in knowledge that
we argue should be acknowledged and analyzed
systematically. In this paper, we synthesize two
themes of recent interest in the ﬁeld of human
geography—literature on illicit geographies and
literature on ambiguity and ignorance—to offer a
methodological intervention for addressing such
Conducting research in areas where illicit activ-
ities are prevalent (or around the illicit, as we will
refer to it) requires wading through ambiguity. Al-
though none of the authors set out to study the illicit
directly, each came to recognize the inﬂuence of illicit
activities on our respective ﬁeld sites and research
topics through encountering “empirical traces”(Frickel,
2014): momentary revelations of something still par-
tially obscured (e.g., observing sudden unexplained
wealth or being warned off certain areas and topics).
These empirical traces offered ﬂeeting glimpses into a
larger landscape of unknowns, shaped our research
practices, and set boundaries on where we went,
uncertainty to the research process (Hilson, 2005;
Nordstrom, 2004). Institutional constraints and
academic conventions conspire to leave such uncer-
tainty mostly ignored (Bhattacharya, 2014;Hicks,
2009). Although the risk of violence and other re-
percussions associated with illicit activities make them
dangerous or difﬁcult for researchers to study directly
e, 2019a), ignoring the inﬂuence of these ac-
tivities is also problematic—both for the conclusions
drawn from the research and in the consequences for
We aim to initiate a collective critical reﬂection on
research and writing practices to address blind spots
and spur more robust research around the illicit,
when illicit activities are not the primary focus but
nevertheless intersect with one’s research. In this
paper, we offer an analytical framework for evalu-
ating ignorances that appear in research around the
illicit through converting missing information into an
ethnographic object (Mair et al., 2012), allowing
researchers to better understand the causes and
consequences of ignorance and account for possible
misinformation. We argue that analyzing knowledge
and ignorance together can be a powerful tool for
gaining insight into the often obscured inﬂuence of
illicit activities on livelihoods, resource use, and
governance practices. Though we focus on ethno-
graphic methodologies in human geography re-
search, an analysis of illicit-generated ignorance is
relevant for researchers from diverse ﬁelds who work
where illicit activities are present.
This paper consists of six sections. Section 2
discusses human geography conceptualizations of
“the illicit”and describes the spheres of ambiguity
that surround illicit activities and the speciﬁc re-
search challenges they pose. Section 3 introduces
concepts from the ﬁeld of agnotology that are useful
for understanding the generation of ignorance in
research around the illicit. Section 4 describes an
Ignorance Assessment Framework for analyzing this
ignorance systematically. In Section 5, we use ex-
amples from our own work to demonstrate the utility
and application of this framework. Section 6 offers
concluding reﬂections on the consequences of ig-
norances that arise in research around the illicit.
II The illicit and the production
1“The illicit”in human geography
The illicit is a consequential but often overlooked
phenomenon in human geography that has gained
increasing attention in numerous subﬁelds. These
include: the geographies of crime and policing
(Evans et al., 2002;Fyfe, 1991;Gilmer, 2014,
2016;Harries, 1971;LeBeau and Leitner, 2011;
Lowman, 1986;Moran and Schliehe, 2017;Peet,
1975;Yarwood, 2007); migration, conﬂict, and se-
curity (Peluso and Watts, 2001;Samers, 2004;Zolberg,
1989); resource frontiers and territorialization (Ballv´
2019b;Kelly and Peluso, 2015;McSweeney et al.,
2018;Watts, 2018); natural resource governance and
conservation (Bocarejo and Ojeda, 2016;Kelly,
2011;Robbins et al., 2009;Woods and Naimark,
2020;Ybarra, 2012); and economic geography on
illicit economies (Brown and Hermann, 2020b;
Gregson and Crang, 2017;Hall, 2010;Hudson,
2020). Although other ﬁelds (sociology, criminol-
ogy, anthropology) have more consistently engaged
with the illicit, geographical theorizations bring out its
socio-spatial and territorial implications. In particular,
geographers have revealed how competing state and
non-state authorities continually negotiate the bounds
between licit and illicit to favor various interests,
discipline populations, and reorganize space (Abraham
and Van Schendel, 2005;Hall, 2013;Hudson, 2014).
In this section, we discuss how the illicit has been
Dev et al. 653
treated in human geography and how we operation-
alize it in our analysis.
Although the term “illicit”is often used as a self-
evident category, it is actually fraught with ambi-
guity. We follow Abraham and Van Schendel (2005),
recognizing that the deﬁning line between licit and
illicit transcends state deﬁnitions of legality. Illicit
activities are pervasive in both social life and state
processes (Abraham and Van Schendel, 2005: 7) and
are integral to global economies and geopolitical
conﬁgurations (Hudson, 2014). Capital, goods, and
actors are not innately illicit or licit, but merely
accumulate these labels as they circulate (Gregson
and Crang, 2017: 212–213). Categorization of the
illicit is shaped by local perceptions of morality,
enforceability, and the threat of violence or other
risks, all of which are highly context-dependent and
dynamic. Certain spaces support a wide range of
unlawful but socially permissible activities and many
illicit and licit activities mutually reinforce each other
(Roitman, 2006). As Hall (2013) describes, “In
spaces that elude the sovereign gaze of the state or
across which it is uncertain or contested, the formal
distinctions between licit and illicitness become
problematic, even meaningless”(373). States thus
extend their authority over new regions and peoples
by dictating the licit/illicit boundary and securing
their exclusive right to enforce it (Vandergeest and
Illicit activities range from complex, highly
adaptable, multinational operations with vast spatial
footprints and vertically integrated production, to
smaller and more localized organizations, to isolated,
unorganized, or small-scale crime performed by in-
dividuals or families as a survival strategy (Hudson,
2014: 779). Illicit organizations are generally ﬂuid;
individual actors or groups may at times merge with
larger organizational structures and later disperse
(Hudson, 2014) or be amorphous in structure to
varying degrees (Finckenauer, 2005). While partici-
pation in these networks may be locally normalized,
persecution by national and/or international entities
encourages a veil of secrecy surrounding the networks’
activities that members often enforce through real or
perceived threats of violence (Andreas, 2011:419–420).
This obscures the entanglements of these networks in a
wide range of licit and illicit economic activities and
social and political affairs (Slack and Campbell, 2016).
Hall (2013) has attributed the paucity of substantive
reﬂections on organized crime in geography to the ex-
tensive scale of illicit networks, the “opacity”around
illicit activities generally, and the ontological uncer-
tainties between licit and illicit activities that make it
difﬁcult to analyze. This paper seeks to address this
Because the impacts of transnational illicit net-
works affect myriad research topics in human ge-
ography (Abraham and Van Schendel, 2005;Hall,
2013;Nordstrom, 2007), we operationalize “the il-
licit”with a focus on illicit commodity production
and trade networks. Though the insights gained are
applicable to illicit activities across the spectrum, the
authors’research sites (Mexico, Peru, Colombia, and
Laos) were heavily affected by patterns of conﬂict
and governance related to the cultivation, processing,
and/or transportation of illicit “drug”crops, con-
trolled by narco-trafﬁcking networks (or “cartels”).
The resource frontiers and borderlands we study have
long been intertwined with the illicit (Castells, 1996:
162; Hall, 2013: 376), making ambiguity around
livelihoods and resource use in these areas particu-
2 Ambiguous spaces: Illicit regimes in
A key set of research questions in human geography
and political ecology intersect in resource frontiers.
Though our research primarily focused on aspects
of these frontiers other than illicit activities, namely
resource management and society-environment
interactions, illicit activities were entangled with
our research topics in the places we work: rural
hinterlands of developing countries; areas less
connected to state services, infrastructure, and
control; and spaces where natural resources and
livelihoods are closely intertwined. This is no co-
incidence. Resource exploitation and illicit networks
are both served by these regions’illegibility to
formal state governance. The ambiguity that per-
meates these places only enhances the messy relations
between ﬁeld-based geographic research, state power,
and illicit activities.
654 Progress in Human Geography 46(2)
Although illicit networks span the globe and af-
fect a variety of spaces, a common set of conditions
conducive to activities like drug crop cultivation,
processing, and transport tend to intersect with
conditions that produce dynamic rural resource
frontiers. Such frontiers are frequently peripheral
regions, especially borderlands where licit market
activity is less supported and state-sanctioned de-
velopment, resource extraction, and land governance
are less established (Bloomer, 2009;Brown and
Hermann, 2020a;Greenough, 2004). They are also
often home to marginalized minority groups and
multiple competing authorities where formal prop-
erty relations are still being institutionalized, liveli-
hoods are changing rapidly, natural resource extraction
is increasing, and state control is being expanded, re-
formulated, and renegotiated (Barney, 2009;Kelly and
Our research sites exemplify how illicit activities
thrive in resource frontiers. In the northern border-
lands of Chihuahua, Mexico, decades of state neglect
and economic hardship have spurred widespread emi-
gration, consolidated agricultural production, and ex-
panded opportunities for a thriving drug trafﬁcking
economy. In the lowlands of the Peruvian Amazon, dense
forest and sheer remoteness have challenged the ex-
pansion of state control and infrastructure while enabling
coca cultivation, land trafﬁcking, and illegal logging to
proceed alongside traditional production systems. Simi-
larly, the rainforests and mangroves of the PaciﬁcCoast
of Colombia shelter revolutionary guerilla armies, para-
militaries, and other violent actors who use cocaine
production and trafﬁcking to fund their causes. In the
highlands of northern Laos, opium cultivation and trade
persist in areas distant from state centers and markets, and
close to increasingly permeable borderlands.
The characteristic distance of resource frontiers
from centers of state control provides a measure of
protection for illicit activities and actors (McSweeney
et al., 2018). Places with contested, incomplete, or
contradictory state oversight are able to serve as hubs
for transnational trafﬁcking of illicit goods (Brown
and Hermann, 2020a). In such spaces, illicit activities
may be irreducibly entangled in social relations of
production, the performance and assertion of resource
claims, and ﬂows of information, capital, and goods.
As a consequence, “confusions between legal and
illegal, public and private, disciplined and wild”can
produce conditions in which rapacious resource ex-
traction and illicit activity both thrive (Tsing, 2003:
e(2019b)even labels frontiers with
thriving illicit regimes as “narco-frontiers”where
“uneven development, internal colonialism, political
violence, and narco-fuelled dispossession”intersect
(211). Brown and Hermann (2020b) refer to land-
scapes of extreme illegibility and little state control as
“black spots." In many of these spaces, illicit networks
or criminal organizations maintain their own form of
order and control, even taking on certain state func-
e, 2019b;Hudson, 2014). Although these
organizations act and work locally, they are made
possible and structured by webs of transnational
connections and distant global processes (Aas, 2007).
Spaces illegible to the state often become targets of
securitization or development interventions (Abraham
and Van Schendel, 2005). Emphasizing the overlap
between frontier spaces and illicit geographies in turn
helps states justify and harness international support
for controversial, even violent, interventions (Ybarra,
2012). Thus, illicit activities can be highly lucrative
but place those who engage in them at risk of punitive
measures from state or transnational authorities,
leading illicit activities to increasingly peripheral
places and increasing secrecy. Because of these dy-
namics, resource frontiers, narco-frontiers, and other
such havens for illicit networks tend to be charac-
terized by pervasive uncertainty, where it is difﬁcult to
ascertain precisely who is doing precisely what. In
such contexts, what limited information is available is
often highly ambiguous.
3 Spheres of ambiguity
The ambiguity surrounding illicit activities has
speciﬁc characteristics that we term spheres of
ambiguity—strategically produced shrouds of
opacity and epistemic uncertainty spanning multiple
dimensions, in which knowledge about certain ac-
tivities, actors, and places remains deliberately oc-
cluded and imprecise, affecting both insiders’and
outsiders’ability to interpret and classify information
to varying degrees. Ambiguity is a powerful analytic,
which Best (2008) deﬁnes as a proclivity for multiple
interpretations that “captures some of the forms of
Dev et al. 655
indeterminacy that exceed and overﬂow the cate-
gories of risk and uncertainty”(355). Spheres of
ambiguity surrounding illicit activities obscure the
timing and geographic extent of illicit activities, the
actors participating, income generated, and the na-
ture of the activities themselves. Strategic ambiguity
is used by these illicit networks as both a tool of
governing and a resistance to state governance (Best,
Ambiguity generates ignorance or incoherence on
a topic through a resistance to a singular interpre-
tation or classiﬁcation (Best, 2012). This is what
McGoey (2007) terms “liminal ignorance”or half-
knowledge. Such a ﬂexibility in interpretation can be
emancipatory in the face of systems that impose false
binaries (Sedgewick, 2008), but can also create
strategic advantages, whether applied to bureaucracy
within organizations (Best, 2012), scientiﬁcﬁndings
(McGoey, 2007;Smithson, 1989), or in the case of
illicit activities, evading the law. According to Best
(2008: 362), ambiguity is reinforced and maintained
by three facets of modern knowledge: the imper-
fection of language with respect to representation;
self-reﬂexivity, which makes knowledge always
subjective and partial; and the historical contingency
of knowledge. Best’s focus on ambiguity centers on
how information is invariably communicated im-
perfectly from one person to another. Ambiguity can
also be cultivated through non-verbal signs, other
context-based cues, and a general atmosphere of
epistemic murk (Taussig, 1984). Within spheres of
ambiguity, even mundane occurrences such as real
estate transactions or ﬁshing locations can become
ambiguous because it is practically impossible (for a
variety of possible reasons) to verify whether such
occurrences are connected to illicit activity. This
ambiguity makes it difﬁcult for a researcher to in-
terpret events with a high degree of certainty.
The considerable risks associated with collecting
data about illicit activities for ﬁeld-based researchers
raise both ethical and methodological questions
(Fitzgerald and Hamilton, 1996). Illicit networks
often have codes of silence enforced throughout the
ranks (e.g., Baker and Faulkner, 1993;Barton and
Davis, 2018;Erickson, 1981;Lindelauf et al., 2009).
Within and surrounding these networks, money and
violence are often used to control the ﬂow of
information, and ambiguity can be an important form
of protection (Ballv´
e, 2019a). Ignorance and ambi-
guity can serve as political tools in that they can
protect against legal incrimination; in a court of law,
ambiguous evidence would be inadmissible or at best
circumstantial by legal standards. Informants may
thus be safer not knowing too much and not revealing
information to others. Due to ever-present threats or
stories of violence, an atmosphere of terror may be
characteristic of ambiguous spaces (Luna, 2018),
with the accompanying epistemic murk described by
Spheres of ambiguity surrounding illicit networks
make the regions and actors that interact with these
networks more difﬁcult to study, particularly for
outside researchers (Ballv´
e, 2019a), even (and perhaps
especially) when illicit activities are not the focus of
the research. Ambiguous signs can include language,
material phenomena, norms and social practices, and
social roles (Best, 2012): 89). Information is often not
fully concealed or revealed to researchers, or even to
their informants, but rather is couched in vague terms
that obfuscate who is involved and what activities,
exactly, are occurring. While ambiguities may be
understood among locals, they tend to be more opaque
for outsiders less accustomed to coded speech and
hidden meanings, and access to illicit spaces can be
e, 2019a). Knowing smiles and other
non-verbal language indicating open secrets among
insiders are common in spaces where both formal and
informal rules and conventions must be negotiated
(Ledeneva, 2011). Actions, too, are ambiguous, when
observed and reported activities may or may not be
connected to illicit materials or transactions. Spheres
of ambiguity can enfold activities that are used to
conceal illicit activities or launder illicit gains, making
it difﬁcult for researchers to parse the relative balance
of income, effort, and inﬂuence from licit versus illicit
activities. Thus, spheres of ambiguity create analytic
challenges that are often ignored, footnoted, or simply
ﬂagged as “gaps”due to the difﬁculty and danger of
investigating these topics more closely.
Despite their resistance to study, documenta-
tion, and quantiﬁcation, illicit networks can have
profound local inﬂuences, adding uncertainty to
otherwise non-controversial research topics like
local resource use or household economics (Devine
656 Progress in Human Geography 46(2)
et al., 2020;McSweeney et al., 2014). According to
...the difﬁculties of deﬁnition and measurement and in
obtaining precise empirical data as to the extent of
illegality should not stand in the way of seeking
to understand its function and the relations between
the legal and illegal. The illegal has clearly become
an integral part of the contemporary phase of capi-
talist development. Put simply, it is too important—
practically and theoretically—to ignore (Hudson,
Though there are promising methods for re-
searching illicit activities in ambiguous or violent
spaces, such as Ballv´
e’s (2019a) investigative eth-
nography, some researchers may have topical, lo-
gistical, or institutional constraints that prevent the
direct study of illicit activity. In either case, the
collected data are likely to have gaps and limitations
that can reproduce ambiguity. We ﬁnd valuable
grounds for addressing these limitations by applying
conceptual tools from agnotology (studies of igno-
rance) to understand the contours, boundaries, and
origins of the spheres of ambiguity encountered in
research around the illicit.
III Knowing what we don’t know: A
Although ignorance produced during research is
often seen negatively, it is also inevitable, and we
argue that studying it can provide useful insights.
Accounting for limitations in research around the
illicit requires understanding the political, social,
and structural mechanisms that actively create ig-
norance, with attention to how absences of certain
data can take on their own meanings (Rappert,
2015). In this section, we introduce concepts
from the science studies ﬁeld of agnotology, which
refers to the study of the social production of ig-
norance or non-knowledge (Proctor and
Schiebinger, 2008) and has been increasingly em-
ployed by geographers. Agnotology is based on the
premise that some ignorances are generated stra-
tegically and that both their causes and inﬂuences
can be studied (McGoey, 2012). Ambiguities and
absences may be used to gain insight into the very
dynamics that create them.
Social theorists in a wide variety of ﬁelds have
attempted to deal with absence (e.g., Bakhtin, 1981;
Bourdieu, 1977;Deleuze and Guattari, 1987;
Foucault, 1970;Luhmann, 1988). Absence and ig-
norance have been topics of concern among a branch
of science studies scholars at least since Kuhn (1962)
argued that scientists actually trade in ignorance by
selectively overlooking ﬁndings that refute the
foundations of their ﬁelds. Dealing with ambiguity
and ignorance has been of growing interest in ge-
ography and other social sciences (Best, 2008;
Croissant, 2018;Frickel, 2014;Galison, 2004;Gross
and McGoey, 2015;Proctor and Schiebinger, 2008;
Smithson, 1989). Amoore (2014), for instance, de-
scribes how scientists calculating national security
risks grapple with predicting ambiguous futures.
These scientists are increasingly held responsible to
“calculate the incalculable”from incomplete data
(Amoore, 2014: 424). As Amoore suggests, the
challenge is not to ﬁnd better data, which may be
impossible, but to ﬁnd better ways of analyzing
incomplete data and accounting for the fallibility of
one’s research methods. Agnotology provides ana-
lytical tools for social scientists to deal with such
ignorance—not as a limitation, but as an object of
social enquiry in itself.
The ﬁeld of agnotology directs researchers to
attend to non-knowledge and how ignorance is
produced, including close examinations of the cul-
tural and political foundations of ignorance (e.g.,
Proctor and Schiebinger, 2008). This is similar to
what Galison (2004) called “antiepistemology”: the
study of the simultaneous production of knowledge
and non-knowledge. By studying ignorance, one is
not trying to illuminate what lies in the darkness, but
rather to assess what is creating the pattern of light
and shadow, how the shadow changes over space and
time, and the consequences of the shadow. Analyzing
patterns of presence and absence simultaneously
allows a more intelligible surface to emerge (e.g.,
Deleuze and Guattari, 1987: 170). The “geography of
ignorance,”or the spatial dimension of ignorance and
knowledge patterns, is important for understanding
how ignorance pervades certain regions, how those
ignorances are produced, and for whom (Proctor,
Dev et al. 657
2008: 26). For example, ignorance about border
regions is produced and reproduced by political and
cartographic conventions, and is further ampliﬁed by
academic disciplinary divisions (Van Schendel,
For McGoey (2012), strategic ignorance can be a
productive force that is more than a mere lack of
knowledge. In other words, ignorance offers its own
type of inﬂuence and information that cannot be
countered simply by adding knowledge. Actively
produced ignorances are those which are made un-
known or unknowable through concealment or
misinformation, whereas passively produced igno-
rances could be known but are not due to inaction,
negligence, or structural constraints (Proctor, 2008).
Much empirical work in agnotology has investigated
the actions of scientists and industry to understand
how public ignorance is socially or politically con-
structed and who beneﬁts. Such ignorance may be
created strategically to deceive others—as self-
protection, in pursuit of ﬁnancial gain, to inﬂu-
ence policy decisions (e.g., Proctor, 2008;Rayner,
2012;Smithson, 1989), or to maintain social and
political order (Gross, 2007). Scientiﬁc conventions
around uncertainty and conditionality can also be re-
sponsible for passively creating non-knowledge (e.g.,
Amoore, 2014;Amrhein et al., 2019;McGoey, 2007).
Most agnotology studies have focused on actively
produced ignorance, but passive ignorance can be
equally important and the twomaybeproducediter-
atively (Durant, 2020). Moreover, actively constructed
ignorances can be disguised as passive (Sanabria,
2016), as is often the case with defensive ignorance
that protects against liability (McGoey, 2007).
Spheres of ambiguity cultivated around illicit
activities generate both active and passive forms of
ignorance. In addition to the strategic ignorance
cultivated by illicit networks, individuals, institu-
tions, media, publics, and researchers all may con-
tribute to producing ambiguities (McGoey, 2012).
Communities around illicit activities can create
passive forms of ignorance through habitual ambi-
guities and avoiding certain topics or places (Luna,
2018). Researchers may generate passive ignorance
through the inability to interpret ambiguous data, or
can actively contribute to the production of ignorance
around the illicit through ethical and safety
considerations that lead them to conceal information
both in the ﬁeld and in writing.
The term “undone science”(Hess, 2015) repre-
sents the practices by which certain types of studies
are conducted and funded while others are left
“undone." In addition to political or economic biases
that inﬂuence these factors, NGOs, state agencies,
researchers, and their funders may avoid potentially
dangerous topics and locations, or populations
considered unreliable or unwelcoming, such as those
associated with certain illicit activities. The conse-
quences of ignorance generated in this way are not
just epistemological, but have material and geopolitical
effects. Biases in where research and funding are di-
rected lead to geographical unevenness in knowledge
andattention(Van Schendel, 2002), which can con-
tribute to unequal distributions of social services or
policy interventions and perpetuate ignorance about
certain issues and social groups.
In sum, ignorance and knowledge are both tools
that can be employed for either liberation or op-
pression (McGoey, 2012). For researchers working
around the illicit, attending to both knowledge and
absences leads to more robust analyses, allowing for
greater consideration of alternative hypotheses,
missing evidence, and incongruities (McGoey, 2012)
and a deeper understanding of how ignorance is gen-
erated and who it serves. Given evidentiary standards in
academia, researchers may be rightfully hesitant to
publish results that include empirical traces and in-
terpreted ambiguities. To bury ambiguous data, how-
ever, presents a reciprocal risk of inaccuracy and has
equivalent material consequences. The Ignorance As-
sessment Framework we present below offers guide-
lines for recognizing and accounting for ignorance in
research around the illicit. While we draw on estab-
lished tools and approaches of agnotology, applying it
self-reﬂexively to one’s own data expands on its typical
IV Analyzing ignorance around the
illicit: A framework
Here we propose an analytic process to account for the
methodological constraints and empirical absences
that arise from researching around the illicit or in other
ambiguous spaces. This Ignorance Assessment
658 Progress in Human Geography 46(2)
ignorances into objects of analysis, allowing the re-
searcher to assess knowledge and ignorance (and
gradations between the two) together. This analytic
process draws attention to gaps in knowledge created
by ambiguity, the known unknowns, which can help
researchers recognize and document what they do
know and responsibly account for what they do not.
By bringing into relief the landscape of knowledge
and ignorance, new patterns and insights emerge.
The Ignorance Assessment Framework consists
of the following steps and guiding questions (see
1. Sensing Absences: How can absences be
2. Describing Ignorance: What are the characteris-
tics of ignorance?
3. Assessing the Causes of Ignorance: How was
These steps constitute an iterative, cyclical pro-
cess described in more detail below with questions
provided as prompts to guide the detection and di-
agnosis of ignorance. In Section 5, we offer speciﬁc
examples of how we applied this process to our own
1 Sensing absences
Recognizing meaningful absences in knowledge
requires reﬂexivity and a willingness to question
one’s data. This step is aimed at detecting ignorances
during ﬁeld work and within the data collected, as
well as identifying the topics in which those igno-
How can ignorance be recognized empirically?
The challenge in detecting empirical absences is that
they defy outright observation. However, absences
may become evident when researchers notice am-
biguities or encounter “empirical traces”: glimpses or
hints of things that are otherwise obscured, though
the meaning may remain uncertain (Frickel, 2014). A
story, warning, or unexpected physical encounter
may provide just enough to reveal one’s ignorance.
Depending on the situation, this could either
Figure 1. Framework for evaluating ignorances in research. Sensing absences, describing ignorance, and assessing the
causes of ignorance are iterative steps that can be done in any order—often simultaneously.
Dev et al. 659
encourage further investigation or warn away from
that line of inquiry. Ambiguous body language such
as knowing glances and smiles may also alert a re-
searcher that a verbal communication may have an
alternate interpretation. In some cases this may be
inaccessible insider knowledge, and in others such
non-verbal cues may be understood by the researcher
but difﬁcult to document or conﬁrm since it was not
said aloud, cannot be asked about, and may directly
contradict what was actually said (Ledeneva, 2011).
Some absences can be recognized more easily when
there are explicit barriers (e.g., when places are “off
limits”or questions “off record”). Empirical igno-
rances will likely be more visible to researchers with
deep local experience and are therefore more likely to
be detected during long-term ethnographic research
than through research methods that require less time
in the ﬁeld engaged in direct personal interaction.
How can ignorance be detected during data
analysis? The ability to detect ignorance in data can
be cultivated through attention to the differences
between what is learned and what could or should be
knowable. Absences may become apparent through
mismatches in triangulation, repeated gaps in reported
information, or by identifying what Star and Bowker
(2007) call “residues," or types of ambiguous infor-
mation that deﬁes categorization and therefore is often
left out or minimized in analyses. Identifying analytic
residues serves as a presencing practice that can give
way to new forms of knowledge (e.g., Puig De la
What areas of knowledge are affected by igno-
rances? Ignorances impact the ability to answer
certain research questions. Identifying topical themes
that likely contain absences and ambiguity aid in
assessing the overall effects of the ignorance on the
study and the research site. These can be assessed by
questioning what types of information may have
been obscured by ambiguity around the illicit as well
as how decision-making in the research process may
have resulted in the avoidance of certain places,
people, or types of questions.
2 Describing ignorance
Describing ignorance ethnographically allows for a
deeper analysis of its effects than just identifying its
existence. This step more ﬁnely delimits the
boundary between what is known and unknown.
What is the form of ignorance? Ignorance can
constitute either a complete lack of information or
ambiguous, erroneous, or misinterpreted informa-
tion. Whereas a researcher often encounters an ab-
sence of information surrounding illicit activities
themselves, ignorance in the form of incorrect or
ambiguous knowledge can also be produced by in-
formants in the process of obscuring the existence of
or participation in illicit activities. This can expand
the sphere of ambiguity well beyond topics directly
related to the illicit.
What is the degree of ignorance? Knowledge and
ignorance form a continuum, not a binary (Frickel,
2014). The degree of ignorance describes the extent
to which something is unknown, on a gradient from
partial to complete ignorance (Rappert, 2015). A
partial ignorance stems from something that is
known of (e.g., via empirical traces) but cannot be
fully seen, understood, interpreted, or quantiﬁed,
often due to some type of ambiguity. Most ignorance
that can be identiﬁed is partial ignorance, in contrast
to a complete absence of information (sometimes
called nescience) in which an “unknown”is entirely
hidden (Gross, 2007).
What are the characteristics of the ignorance? We
adapt Croissant’s (2014,2018) typology to describe
ignorance by adding relationality,spatiality, and
These additional characteristics are
relevant for understanding ignorance geographically
in a ﬁeld-based context (rather than a textual or
laboratory context). Thus, ignorance can be assessed
with respect to the following characteristics:
Chronicity: The time scale of ignorance; non-
knowledge can occur in the present, can be prospec-
tive (in the future), or retrospective (occurring in the
Granularity: The “texture”of the non-knowledge; the
size of detail missing, from minute details to large
swaths of information. Often related to spatiality.
Scale: The structural level assessed in the analysis of
ignorance; relates to, but operates independently of,
granularity. This can range from individual decision-
660 Progress in Human Geography 46(2)
making or cognitive processes to cultural formations or
ideologies that shape broad patterns.
Intentionality: The intentions behind missing informa-
tion. Ranges from direct intent (lying or misinformation),
to concealment and nondisclosure, to unintentional ig-
norance or uncertainty from the informant or researcher.
Ignorance around the illicit is usually intentional by
someone, though the relation may be indirect.
Relationality: Refers to how the one who is ignorant
relates to the object of ignorance. Reﬂects an identiﬁcation
of who is ignorant and how that ignorance travels socially.
Spatiality: The geographical dimensions of ignorance
and knowledge; reﬂects how certain locations may be
more or less known or legible, or particular areas more
or less accessible to research in general.
Materiality: Material characteristics of the objects of
ignorance (data, images, people, services, events, money,
documents, etc.) and how they exist in the world.
3 Assessing the causes of ignorance
Interacting with spheres of ambiguity surrounding
the illicit, ignorance can be reinforced at multiple
points in the research process. The “ignorance supply
chain”is made up of the nested and interdependent
institutions and processes from the inception of a
study to the dissemination of results. Ignorance can
be actively or passively created at each stage, with
further impacts downstream.
How is ignorance generated in situ? Informants,
collaborators, state actors, and institutions can all
generate ignorance. Assessing the production of
ignorance requires identifying the actors, prac-
tices, and structures that produce, maintain, and
disseminate various forms of ignorance. Are ig-
norances generated actively (e.g., concealment or
deceit) or passively (e.g., avoidance or neglect)? Is
ignorance the result of individual actions or are
there institutional structures or constraints that
promote ignorance? How is secrecy enforced and
maintained? In the case of research around the
illicit, these could range from threats of violence
(to the researcher or to their informants), to
principled aversion (e.g., where illicit activity is
considered morally distasteful and thus not dis-
cussed), to cultural norms and habits that generate
ambiguity. The high stakes and overlapping factors
driving secrecy around illicit activities and net-
works make these distinctions particularly difﬁcult
How do research practices reinforce ignorance?
Researchers are embedded in an institutional
context wherein funders, advisers, collaborators,
review boards, stakeholders, and publishers can
all contribute to the ignorance supply chain by
steering the researcher away from certain topics,
populations, and places. This chain has corollaries
variety of actors across scales and levels of au-
thority contribute to the production of ignorance.
Thus, ignorance is created, fostered, tolerated, and
reproduced both through research practices and by
processes external to them, and often these are
entangled. Interrogating how our own research
practices engage or ignore the presence of the
illicit brings attention to the ways we “presence”
and “absence”certain details (Rappert, 2015),
which are sometimes political or ethical in nature.
Gervais et al. (2001: 424) describe four interre-
lated categories of absences produced during re-
search, and we have added positional absences to
Theoretical: a result of inadequate theorization or
inability to conceptualize.
Methodological: due to insufﬁcient or inappropriate
methods, inadequate sampling, or other methodological
constraints for presencing details.
Empirical: resulting from concealment, censoring,
misinformation, or ambiguity in the empirical details.
Analytical: when absences are naturalized, ignored,
left out, or are insufﬁciently interpreted.
Positional: when information is excluded from re-
porting for ethical, social, or political considerations.
Assessing how ignorances are generated at each
stage of the research process helps the researcher
hone their practices, as well as account for their
Dev et al. 661
V Evaluating illicit-related ignorances
in research on rural livelihoods
Using examples from our own work, we demonstrate
how the Ignorance Assessment Framework outlined
above can be applied to reveal the effects of the illicit
on livelihoods, resources, and governance. This is best
seen as an effort of guided reﬂexivity—systematically
evaluating how the presence of the illicit interacted
with our research practices and impacted our ability
1 Sensing absence: Recognizing the shadows
through empirical traces
The presence of illicit activities in our ﬁeld sites was
revealed through empirical traces that provided only
ambiguous details about the activity or who partic-
ipated. Certain places were unexpectedly off limits,
certain questions repeatedly unanswered; we stum-
bled upon evidence of illicit activities (hearing about
packages of drugs washed up on the beach or being
led through a ﬁeld of coca during an interview);
interlocutors made certain comments only off record.
Sometimes the illicit showed up in storytelling that
blurred fact and ﬁction, or emerged as coded or
ambiguous language (e.g., ﬁshing for “white ﬁsh”—
smuggling cocaine by boat). Other times, references
to illicit activities were included in broader narratives
but rarely seen ﬁrsthand (e.g., rumors that certain
villages or ethnic groups cultivated opium or coca).
Despite the ambiguity and secrecy surrounding these
encounters, taking note of them enhanced under-
standing of our regions of study. The accumulation of
empirical traces elucidated gaps in our knowledge
and guided more careful, systematic consideration
of how these ignorances affected our research
Through the Ignorance Assessment Framework,
we identiﬁed three themes most likely impacted by
absent, ambiguous, and erroneous information re-
lated to the illicit in our cases: (a) livelihoods, labor
allocation, and income; (b) natural resource use, land
use, and land transfers; and (c) governance, social
relations, and power dynamics (Table 1). The fol-
lowing section is organized around these themes.
2 Describing and characterizing ignorance
(a) Livelihoods, labor allocation, and income. Although
we knew that illicit activity occurred in our re-
spective ﬁeld sites, informants rarely reported par-
ticipating in or beneﬁting from illicit activities
themselves. Instead, they mentioned others engaging
in such activities, or referenced them occurring in the
community in broad strokes. The resulting ignorance
can be described in terms of its granularity—the
researcher has some coarse knowledge (that illicit
activities exist) but is missing ﬁner-grained speciﬁcs
about the activity and its importance (who partici-
pates and how much they earn). For example, we may
know that cannabis is grown by some households
(coarse-grained knowledge), but not which house-
holds or how much they are growing (ﬁne-grained
ignorance). This constrains the analysis of economic
activity in the community. Alternatively, a researcher
may pick up ﬁne-grained detail, such as knowledge
that a particular individual is involved in an illicit
activity, but remain uncertain about the scale, or how
widespread that activity is.
In surveys or interviews, the time spent or income
gained through illicit livelihoods may be completely
omitted (absence) or ascribed to licit activities in-
stead (misinformation). For example, a ﬁsher in
coastal Colombia who earns additional income using
their boat to supply drug trafﬁckers with gasoline and
food may attribute that time spent and income earned
to ﬁshing. This obscures data about the illicit live-
lihood and creates ignorance in the form of misin-
formation about ﬁshing livelihoods. If the income
earned from illicit activities is omitted entirely, it
results in underreported household income. This can
be problematic when analyzing how households are
making labor decisions or whether traditional live-
lihoods remain viable. This type of gap may be
identiﬁed through triangulating or identifying mis-
matches in data, or may only be suspected. Never-
theless, recognizing where ignorances likely occur in
the data aids in identifying limitations and better
interpreting the data despite them. For example, we
can recognize that a household or community’s
surprising—if irregular—material wealth must be
attributed to more than just sales of ﬁsh or non-timber
forest products, even if we cannot account for how
662 Progress in Human Geography 46(2)
much of it is generated by undescribed activities or
be certain that they are illicit.
(b) Natural resource use, land use, and land transfers. Land
and natural resource use are also shaped by illicit ac-
tivities, though the drivers of change are often obscured
(Tellman et al., 2020). The capital from illicit trade could
be used to purchase land for ranches or plantations, scale
up ﬁshing and agricultural production, or invest in a
sawmill, which can all be used as fronts to launder illicit
earnings. These hidden but consequential links between
illicit activities and shifts in land use and resource ex-
traction are common across our research sites.
Illicit activities can obscure data about natural
resource use, such as harvest pressure or extraction
rates. In the Peruvian Amazon, most logging occurs
in unpermitted areas, though most timber circulating
has certiﬁcates claiming legal origin (Sears and
Pinedo–Vasquez, 2011). If a timber processor
claims their logs were harvested legally, the in-
tentionality behind this misinformation remains
unclear—whether the processor is actually unaware
Table 1. Assessing the effects of illicit-related ignorance.
(a) Livelihoods, labor allocation, and income
Research themes affected Knowledge obscured Possible misinterpretations
Viability and importance of livelihood
Time allocation of labor
Relative contributions to household
extensiﬁcation, or migration
Participation, income, and
importance of illicit activities
Decisions about traditional
livelihoods and labor activities
Reasons individuals emigrate or not
(e.g., danger or supplemental
Trade and trafﬁcking quantities and
Overreported licit livelihood activities
Extra income or apparent wealth ascribed
to licit livelihoods
Total income underreported
Decision-making about livelihoods
(b) Natural resource use, land use, and land transfers
Research themes affected Knowledge obscured Possible misinterpretations
Natural resource use and importance
Drivers of land use change
Decision-making around land use
Land ownership and access
Harvest pressure of forest products
or production quantity of crops
Details of land sales and transfers
(e.g., source of income,
Land use activities
Impacts of trafﬁcking on access
Overestimation or underestimation of
illicit crop production or harvest of
Land-use change misattributed
Resource use and ecological impacts
Overestimation of access and mobility if
dangers are unaccounted for
(c) Governance, social relations, and power dynamics
Research themes affected Knowledge obscured Possible misinterpretations
Loci of power
Social norms, alliances, and
Drivers of local governance
How power is exercised, gained, and
How relations and alliances are
Roles of important actors
Drivers of decisions misattributed due to
rumors and propaganda
Power of ofﬁcial ofﬁces/policies
Role of illicit regimes underrepresented
Economic drivers overemphasized (over
Dev et al. 663
of the timber’s origins, or is actively concealing the
illicit origin of the timber. Although there are many
factors that can result in absence or error (e.g., recall
bias), the ignorances associated with the illicit are
usually intentional, though the relationality of the
ignorance (to whom it pertains and from where it is
generated), as in the above scenario, may be difﬁcult
to parse. An informant may intentionally contribute
to a researcher’s ignorance (e.g., refusing interviews,
concealing information, or giving misinformation) or
may be poorly informed themselves. Ambiguity in
relationality and intentionality is prevalent as it
serves protective functions, and considering these
limitations to certainty may help the researcher
navigate the larger landscape of presences and ab-
sences in their work.
Land transfers tend to be poorly documented in
remote areas, and it can be nearly impossible to
ascertain whether illicit activities fund or motivate a
land purchase (Tellman et al., 2020). During research
in northern Mexico, for example, many interviewees
refused to answer interview questions about ranch
sales, leaving gaps within the project’smaterial
interview data. Others openly recounted past in-
stances where ranches were purchased by cartel
members for growing or transporting drugs but rarely
discussed recent examples of this, leaving it unclear
whether the practice had stopped or information was
being concealed. Regardless, the lapse of time ren-
dered information about past illicit activity safer to
share, creating temporal biases in knowledge. Thus,
the chronicity of ignorance was characterized by an
absence of information in the present tense.
(c) Governance, social relations, and power dynam
ics. Ambiguity around the illicit challenges re-
searchers’understanding of decision-making in the
socio-political sphere at multiple scales—from in-
dividuals and households to community and re-
gional governance bodies. For example, the use of
land for particular ends, such as alluvial gold mining
or forest clearcutting, may be permitted by local
governance institutions only because illicit net-
works threaten or bribe local leaders. As outsider
researchers, it is difﬁcult to see what drives the
decisions of a governance body, and these mo-
tives may only be revealed through legal actions
(e.g., if a leader is convicted or accused of ac-
cepting bribes) or acts of violence (e.g., if leaders
are threatened or assassinated).
Illicit activities carry symbolic power that can
affect resource governance broadly. It is common for
accusations of illicit activity on one hand, or threats
of violence on the other, to be mobilized to legitimize
or enforce claims to land and resources. Such public
claims and narratives can obscure and conceal why
policy and governance decisions are being made. For
example, in Laos, investors and state authorities
categorize certain ethnic groups and landscapes as
sources of opium production, though exact locations
of opium ﬁelds are rarely known or monitored. This
has been used to justify both the exclusion from and
targeting of those landscapes and ethnic groups for
agribusiness concessions (Dwyer, 2014;Lu, 2017).
Thus alleged associations with illicit activity can
shape broader geographies of ignorance at multiple
3 Tracing the origins and drivers of ignorance
Tracing the origins of ignorance helped us recognize
the ways our research practices produce and replicate
ignorance. Spheres of ambiguity around illicit ac-
tivities interact with researchers’practices and
decision-making as well as academic or institutional
conventions to contribute to the ignorance supply
chain (Tab le 2).
Conceptualizations of research questions and
study designs reﬂect and are constrained by gaps in
the literature. For some of us, the paucity of literature
discussing the importance of the illicit for our re-
search topics limited our ability to plan for en-
counters or anticipate their effects. In other cases, we
were broadly aware of illicit activities in potential
research sites and deliberately chose to avoid certain
areas and topics out of safety concerns for ourselves,
research assistants, and informants. In some cases the
individuals or institutions overseeing our work, in-
cluding funders and Institutional Review Boards,
restricted where we could go, what we could ask, or
what methods we could use. Methodological ab-
sences arose when we were instructed not to conduct
certain types of interviews or use certain methods for
collecting data. Though often unintentional, well-
664 Progress in Human Geography 46(2)
justiﬁed, or out of our control, these decisions and
restrictions became “absenting”practices within the
research process. Normalization of such avoidance
can further perpetuate theoretical absences or under-
theorization of topics important to these places.
During data collection, maintaining research ac-
cess depended on respecting local social norms, and
was inﬂuenced by danger and risk. At times, this
meant avoiding certain spaces and people, averting
one’s eyes, or pretending not to witness certain ac-
tivities. Navigating this terrain limited the informa-
tion that we could collect, contributing to the
existence and persistence of empirical ignorances.
The ignorances and avoidance practices of research
collaborators, NGOs, or government agencies also
contributed to empirical ignorance. NGOs, for ex-
ample, often rely on remaining “apolitical," avoiding
involvement in political conﬂicts or crossing local
authorities by withholding information or remaining
During analysis and writing, decisions on how to
treat empirical ignorances around the illicit can
transfer those ignorances to wider audiences. Ana-
lytical ignorances are perpetuated when a researcher
does not include ambiguous data in their analysis.
Without methodically assessing ignorance, it is easy
to minimize or ignore data that is anecdotal or based
on incomplete knowledge, potentially leading to
further misinterpretations (see Table 1). At the dis-
semination stage, concerns about data resolution,
conﬁdence, and ethics shape how individuals and
research teams navigate these decisions. Researchers
may generate positional ignorance through their need
to maintain conﬁdentiality and secrecy on behalf of
Table 2. Ignorance Supply Chain—locating causes of ignorance within the research process.
Research stage Potential drivers of ignorance Examples from the authors
Study design Limited planning due to lack of information in the
Avoiding locations or topics due to logistical or
Methods and scope restricted due to safety and
ethical concerns or political commitments (by
IRB, host country, funders, or research
Impact of trafﬁcking on small-scale ﬁshing absent
in the literature, thus not considered relevant.
Warnings of danger and a recent murder
deterred the investigation of illegal logging;
study sites were chosen to avoid violence.
Funders rejected research proposals due to
concern about danger in the region; IRB
approval was contingent on assurance not to
ask questions about illicit activities.
Data collection Censorship of activities, contacts, and questions
due to safety concerns and local norms.
Information omitted, concealed, obstructed, or
denied; spheres of ambiguity built into the
Political barriers to accessing reliable secondary
data (from public sources, partnerships, or
Vehicles perceived to be owned by illicit actors
Machinery in an illicit goldmine
explicitly concealed by informants who feared
retaliation from superiors should the
machines be destroyed later.
Illicit incomes seldom reported in public data;
remote sensing images do not include
classiﬁcation for illicit crops; government data
on illicit crop eradication not public.
Details excluded to protect conﬁdentiality and
safety of participants and communities.
Data excluded from analysis due to poor data
quality, conﬁdence, or resolution.
Disciplinary norms or analytical standards
discount the value of partial knowledge.
Farmers reverting to illicit crops from failed
substitution crops went unreported to
protect the community from eradication
UNODC data on illicit activities too highly
aggregated to clarify uncertainties from
Avoided publishing about illicit activities to
ensure ongoing access to ﬁeld sites.
Dev et al. 665
their informants’and themselves. Researchers may
choose not to write about illicit activity for fear of
exposing certain regions or populations to scrutiny or
to protect their own safety, relationships, and access
to research sites. The social and political sensitivity
of reporting on the illicit, even tangentially, can lead
researchers to avoid the topic rather than risk un-
wanted attention or misuse of their results, though
maintaining these ignorances in some cases may
support ongoing inequities and injustices.
Illicit activities have impacts throughout the socio-
natural world that are of interest to scholars across
human geography and other disciplines. This is es-
pecially true in certain frontier regions where in-
creasing transnational connectivity has intensiﬁed
illicit crop production, smuggling, and conﬂict
(Abraham and Van Schendel, 2005) and where illicit
activities are entangled with regional economies
(Hudson, 2014). Thus, research in such regions is
especially imbricated with the illicit, from conser-
vation scientists studying deforestation associated
with illicit crops (e.g., Bradley and Millington, 2008)
to anthropologists working with marginalized pop-
ulations in conﬂict zones (e.g., Luna, 2018). Though
illicit activities create their own geographies (Hall,
2013), spheres of ambiguity around illicit activities
are liable to obscure the factors that drive decision-
making, environmental changes, and socio-spatial
patterns and make interpretation difﬁcult. Why are
people migrating? Why are they selling land? What
drives investments? Answers may vary wildly
(McSweeney et al., 2017), and common logics may
be elusive for outsider researchers. The proposed
Ignorance Assessment Framework can attune re-
searchers to ambiguities, help them interrogate their
own ignorance, and guide research and reporting.
Ignorance around the illicit has material conse-
quences, often making development and conserva-
tion efforts in drug trafﬁcking regions out of touch
and counterproductive. Alternative development
projects and crime mitigation policy are particularly
prone to perverse outcomes or failure when it is
poorly understood how illicit activities shape ev-
eryday decisions (e.g., D´
avalos et al., 2016;Gilmer,
2014;Salisbury and Fagan, 2013;Sears and Pinedo–
Vasquez, 2011). Similarly, conservation efforts are
vulnerable to environmental damage (e.g., “narco-
deforestation”) and local conﬂict when cartels and
paramilitaries establish operations in protected areas
due to the relative lack of population and surveillance
(e.g., Devine et al., 2020;Greenough, 2004;Kelly
and Ybarra, 2016;McSweeney et al., 2014;Sesnie
et al., 2017;Wrathall et al., 2020;Ybarra, 2012).
Meanwhile, the communities living in resource
frontiers and other ambiguous spaces become less
legible to the world. These regions are already subject
to mischaracterizations (e.g., marginal, criminal, and
chaotic) by powerful actors who stand to gain from
certain narratives and thus strategically cultivate
ambiguity: from state authorities implementing terri-
torializing projects or NGOs pursuing alternative
development plans to corporations seeking to proﬁt
from natural resource extraction or illicit networks
attempting to evade the state’s gaze. This speaks to
Best’s(2008)assertion that ambiguity can present at
once, “an object of governance—that one seeks to
contain—a strategic asset in governance—that is
exploited to achieve a certain end—and a limit to
governance—as that which exceeds efforts to control
it”(Best, 2008: 363). Although this ambiguity poses
challenges to researchers, to leave such spaces un-
studied is to surrender control of the regional narrative
to those in power. It is important to recognize that
while empirical traces and sensed absences are in-
sufﬁcient to deﬁnitively prove the role of the illicit,
limiting analyses to unambiguous data and licit ac-
tivities is likely to misrepresent or obscure the shadow
economies and power structures that impact the lives
of millions of people around the globe.
While the epistemological constraints of ﬁeld
research are hardly reducible to the presence of the
illicit, we have found it to require special consider-
ations. We suggest that spheres of ambiguity sur-
rounding illicit activities generate patterns of
constructed ignorance that have yet to be framed within
the literature. Spheres of ambiguity impact each step in
the research process, contributing to an ignorance
supply chain that limits what can be known. Re-
searchers working around the illicit must contend with
how to evaluate the knowledge-ignorance landscape
that is generated. We drew on agnotology to construct
666 Progress in Human Geography 46(2)
an Ignorance Assessment Framework that we offer to
scholars in human geography and beyond to more
explicitly grapple with the causes and implications of
spheres of ambiguity and the subsequent ignorance
generated around the illicit. We suggest that ignorance
related to illicit activity deserves better acknowledge-
ment and scrutiny due to its pervasive impacts not only
on research outcomes, but on regional and global ge-
ographies, interventions, and the lives of local people.
This paper is the product of a joyful and highly collabo-
rative engagement spanning nearly four years among the
authors, with extensive inputs from a few notable parties.
We want to thank Progress in Human Geography’s editor,
Louise Amoore, and our two anonymous reviewers for
their careful but challenging and constructive engagement
with the piece. The ideas presented here have also beneﬁted
greatly from astute insights from Patrick Baur, who pro-
vided feedback on an early version of this manuscript. Our
foundational concepts were reﬁned through a discussion
with Brittany Gilmer, who participated in our panel,
“Encounters with the Illicit: Characterizing Landscapes of
Ignorance”at the Annual AAG meeting in 2019 in
Washington, D.C.; dialogue among the panelists and those
in attendance helped inspire and guide our initial framing
of this paper.
Declaration of conﬂicting interests
The author(s) declared no potential conﬂicts of interest
with respect to the research, authorship, and/or publication
of this article.
The author(s) received no ﬁnancial support for the re-
search, authorship, and/or publication of this article.
1. We excluded Croissant’sontological/epistemological
category, as our framework concerns a subset of
epistemological ignorances rather than cognitive limi-
tations or stochastic/probabilistic uncertainties.
Aas KF (2007) Globalization & crime. In: Key Approaches to
Criminology. Los Angeles, London: SAGE Publications.
Abraham I and van Schendel W (2005) Introduction: the
making of illicitness. In: van Schendel W and
Abraham I (eds), Illicit Flows and Criminal Things:
States, Borders, and the Other Side of Globalization.
Bloomington, IN: Indiana University Press, pp. 1–37.
Amoore L (2014) Security and the incalculable. Security Dia-
logue 45(5): 423–439. DOI: 10.1177/0967010614539719.
Amrhein V, Greenland S and McShane B (2019) Scientists
rise up against statistical signiﬁcance. Nature 567(7748):
305–307. DOI: 10.1038/d41586-019-00857-9.
Andreas P (2011) Illicit globalization: myths, miscon-
ceptions, and historical lessons. Political Science
Quarterly 126(3): 403–425.
Baker WE and Faulkner RR (1993) The social organization
of conspiracy: illegal networks in the heavy electrical
equipment industry. American Sociological Review
58(6): 837–860. DOI: 10.2307/2095954.
Bakhtin MM (1981). In: Holquist M, Emerson C and
Holquist M (eds), The Dialogic Imagination: Four
Essays. Austin: University of Texas Press.
e T (2019a) Investigative ethnography: a spatial
approach to economies of violence. Geographical
Review 110: 238–251. DOI: 10.1111/gere.12347.
e T (2019b) Narco-frontiers: a spatial framework for
drug-fuelled accumulation. Journal of Agrarian Change
19(2): 211–224. DOI: 10.1111/joac.12300.
Barney K (2009) Laos and the making of a ’relational’
resource frontier. Geographical Journal 175(2):
146–159. DOI: 10.1111/j.1475-4959.2009.00323.x.
Barton A and Davis H (eds), (2018) Ignorance, Power and
Harm - Agnotology and the Criminological Imagi-
nation. In: Critical Criminological Perspectives.
London: Palgrave Macmillan.
Best J (2008) Ambiguity, Uncertainty, and Risk: Re-
thinking Indeterminacy1. International Political So-
ciology 2(4): 355–374. DOI: 10.1111/j.1749-5687.
Best J (2012) Bureaucratic ambiguity. Economy and So-
ciety 41(1): 84–106. DOI: 10.1080/03085147.2011.
Bhattacharya S (2014) Institutional review board and in-
ternational ﬁeld research in conﬂict zones. PS:
Dev et al. 667
Political Science & Politics 47(4): 840–844. DOI: 10.
Bloomer J (2009) Using a political ecology framework to
examine extra-legal livelihood strategies: a Lesotho-
based case study of cultivation of and trade in can-
nabis. Journal of Political Ecology 16(1): 49–69.
Bocarejo D and Ojeda D (2016) Violence and conservation:
beyond unintended consequences and unfortunate
coincidences. Geoforum 69: 176–183. DOI: 10.1016/j.
Bourdieu P (1977) Outline of a Theory of Practice.
Cambridge: Cambridge University Press.
Bradley AV and Millington AC (2008) Coca and colonists:
quantifying and explaining forest clearance under
coca and anti-narcotics policy regimes. Ecology and
Society 13(1): 31. DOI: 10.5751/ES-02435-130131.
Brown SS and Hermann MG (2020a) Introduction: Setting
the Stage. In: Brown SS and Hermann MG (eds),
Transnational Crime and Black Spots: Rethinking
Sovereignty and the Global Economy. International
Political Economy Series. London: Palgrave Macmillan
UK, pp. 1–12. DOI: 10.1057/978-1-137-49670-6_1.
Brown SS and Hermann MG (eds), (2020b) Transnational
Crime and Black Spots: Rethinking Sovereignty and
the Global Economy. In: International Political
Economy Series. London: Palgrave Macmillan UK.
Castells M (1996) The Rise of the Network Society. In: The
Information Age: Economy, Society and Culture
Volume 1. Malden, Mass: Blackwell Pub.
Croissant JL (2014) Agnotology: Ignorance and Absence
or Towards a Sociology of Things That Aren’t There.
Social Epistemology 28(1): 4–25. DOI: 10.1080/
Croissant JL (2018) Agnotology: Ignorance and Absence, or
Towards a Sociology of Things that Aren’t There. In:
Meusburger P, Heffernan M and Suarsana L (eds),
Geographies of the University. Knowledge and Space.
Cham: Springer International Publishing, pp. 329–351.
avalos LM, Sanchez KM and Armenteras D (2016) De-
forestation and coca cultivation rooted in twentieth-
century development projects. BioScience Oxford Ac-
ademic 66(11): 974–982. DOI: 10.1093/biosci/biw118.
Deleuze G and Guattari F (1987) A Thousand Plateaus:
Capitalism and Schizophrenia (Tran. B Massumi).
2nd ed. Minneapolis: University of Minnesota Press.
Devine JA, Wrathall D, Currit N, et al. (2020) Narco-
Cattle Ranching in Political Forests. Antipode 52(4):
1018–1038. DOI: 10.1111/anti.12469.
Durant JL (2020) Ignorance loops: how non-knowledge
about bee-toxic agrochemicals is iteratively produced.
Social Studies of Science 50(5): 751–777. DOI: 10.
Dwyer MB (2014) Micro-geopolitics: capitalising security in
laos’s golden quadrangle. Geopolitics. Routledge 19(2):
377–405. DOI: 10.1080/14650045.2013.780033.
Erickson BH (1981) Secret societies and social structure. Social
Forces 60(1): 188–210. DOI: 10.1093/sf/60.1.188.
Evans D, Fyfe N and Herbert D (2002) Crime, Policing and
Place: Essays in Environmental Criminology. Routledge.
Finckenauer JO (2005) Problems of deﬁnition: what is
organized crime? Trends in Organized Crime 8(3):
Fitzgerald J and Hamilton M (1996) The consequences of
knowing: Ethical and legal liabilities in illicit drug re-
search. Social Science & Medicine 43(11): 1591–1600.
Foucault M (1970) The Order of Things: An Archaeology
of the Human Sciences. New York: Vintage Books.
Frickel S (2014) Absences: methodological note about
nothing, in particular. Social Epistemology 28(1):
86–95. DOI: 10.1080/02691728.2013.862881.
Fyfe NR (1991) The police, space and society : the geog-
raphy of policing. Progress in Human Geography
15(3): 249–267. DOI: 10.1177/030913259101500301.
Galison P (2004) Removing knowledge. Critical Inquiry
31(1): 229–243. DOI: 10.1086/427309.
Gervais MC, Morant N and Penn G (2001) Making sense
of “absence”: towards a typology of absence in social
representations theory and research. Journal for the
Theory of Social Behaviour 29(4): 419–444. DOI: 10.
Gilmer B (2014) Political Geographies of Piracy: Con-
structing Threats and Containing Bodies in Somalia.
Gilmer B (2016) Fishermen, pirates, and the politics of aid:
an analysis of the Somali Fishermen Registration
Programme. Geoforum 77: 106–113. DOI: 10.1016/j.
668 Progress in Human Geography 46(2)
Greenough P (2004) Bio-ironies of the Fractured Forest:.
In: Slater C (ed), In Search of the Rain Forest. Duke
University Press, pp. 167–203. DOI: 10.1215/
Gregson N and Crang M (2017) Illicit economies: cus-
tomary illegality, moral economies and circulation.
Transactions of the Institute of British Geographers
42(2): 206–219. DOI: 10.1111/tran.12158.
Gross M (2007) The unknown in process. Current Sociology
55(5): 742–759. DOI: 10.1177/0011392107079928.
Gross M and McGoey L (2015) Routledge International
Handbook of Ignorance Studies. Routledge.
Hall T (2010) Economic geography and organized crime: a
critical review. Geoforum 41(6): 841–845. DOI: 10.
Hall T (2013) Geographies of the illicit. Progress in Hu-
man Geography 37(3): 366–385. DOI: 10.1177/
Harries KD (1971) The Geography of American Crime,
1968. Journal of Geography 70(4): 204–213. Rout-
ledge. DOI: 10.1080/00221347108981621.
Hess DJ (2015) Undone science and social movements: a
review and typology. In: Gross M and McGoey L
(eds), Routledge International Handbook of Igno-
rance Studies. London and New York: Routledge.
Hicks D (2009) Evolving regimes of multi-university re-
search evaluation. Higher Education 57(4): 393–404.
Hilson G (2005) Strengthening artisanal mining research
and policy through baseline census activities. Natural
Resources Forum 29(2): 144–153. DOI: 10.1111/j.
Hudson R (2014) Thinking through the relationships be-
tween legal and illegal activities and economies:
spaces, ﬂows and pathways. Journal of Economic
Geography 14(4): 775–795. DOI: 10.1093/jeg/lbt017.
Hudson R (2020) The illegal, the illicit and new geogra-
phies of uneven development. Territory, Politics,
Governance 8(2): 161–176. Routledge. DOI: 10.
Kelly AB (2011) Conservation practice as primitive ac-
cumulation. Journal of Peasant Studies 38(4):
683–701. DOI: 10.1080/03066150.2011.607695.
Kelly AB and Peluso NL (2015) Frontiers of commodi-
ﬁcation: state lands and their formalization. Society &
Natural Resources 28(5): 473–495. DOI: 10.1080/
Kelly AB and Ybarra M (2016) Introduction to themed
issue: “Green security in protected areas”.Geoforum
69: 171–175. DOI: 10.1016/j.geoforum.2015.09.013.
Kuhn T (1962) The Structure of Scientiﬁc Revolutions.
Chicago: University of Chicago Press.
LeBeau JL and Leitner M (2011) Introduction: progress in
research on the geography of crime. The Professional
Geographer 63(2): 161–173. Routledge. DOI: 10.
Ledeneva A (2011) Open secrets and knowing smiles. East
European Politics and Societies: And Cultures 25:
720–736. SAGE Publications Inc. DOI: 10.1177/
Lindelauf R, Borm P and Hamers H (2009) The inﬂuence
of secrecy on the communication structure of covert
networks. Social Networks 31(2): 126–137. DOI: 10.
Lowman J (1986) Conceptual issues in the geography of
crime: toward a geography of social control. Annals of
the Association of American Geographers 76(1):
81–94. DOI: 10.1111/j.1467-8306.1986.tb00105.x.
Lu JN (2017) Tapping into rubber: China’s opium re-
placement program and rubber production in Laos.
The Journal of Peasant Studies 44(4): 726–747.
Routledge. DOI: 10.1080/03066150.2017.1314268.
Luhmann N (1998) The ecology of ignorance. In: Luh-
mann N (ed), Observations on Modernity. Stanford,
CA: Stanford University Press.
Luna S (2018) Affective atmospheres of terror on the
mexico-u.s. border: rumors of violence in reynosa’s
prostitution zone. Cultural Anthropology 33(11):
58–84. DOI: 10.14506/ca33.1.03.
Mair J, Kelly A and High C (2012) Introduction: making
ignorance an ethnographic object. In: High C, Kelly A
and Mair J (eds). The Anthropology of Ignorance: An
Ethnographic Approach. New York: Palgrave Mac-
McGoey L (2007) On the will to ignorance in bureaucracy.
Economy and Society 36(2): 212–235. DOI: 10.1080/
McGoey L (2012) Strategic unknowns: towards a soci-
ology of ignorance. Economy and Society 41(1):
1–16. DOI: 10.1080/03085147.2011.637330.
McSweeney K, Nielsen EA, Taylor MJ, et al. (2014) Drug
policy as conservation policy: narco-deforestation. Scien-
ceAmerican Association for the Advancement of Science
343(6170): 489–490. DOI: 10.1126/science.1244082.
Dev et al. 669
McSweeney K, Richani N, Pearson Z, et al. (2017) Why do
narcos invest in rural land? Journal of Latin American
Geography 16(2): 3–29. DOI: 10.1353/lag.2017.0019.
McSweeney K, Wrathall DJ, Nielsen EA, et al. (2018)
Grounding trafﬁc: the cocaine commodity chain and
land grabbing in eastern Honduras. Geoforum 95:
122–132. DOI: 10.1016/j.geoforum.2018.07.008.
Moran D and Schliehe AK (2017) Carceral Spatiality: Di-
alogues between Geography and Criminology. Springer.
Nordstrom C (2004) Shadows of war: violence, power, and
international proﬁteering in the twenty-ﬁrst century.
In: California Series in Public Anthropology. Ber-
keley: University of California Press.
Nordstrom C (2007) Global Outlaws: Crime, Money, and
Power in the Contemporary World. University of
Peet R (1975) The geography of crime: a political critique∗.
The Professional Geographer 27(3): 277–280. Rout-
ledge. DOI: 10.1111/j.0033-0124.1975.00277.x.
Pelus NL and Watts M (2001) Violent Environments.
Cornell University Press.
Proctor RN (2008) Agnotology: a missing term to describe
the cultural production of ignorance (and its study).
In: Proctor RN and Schiebinger L (eds), Agnotology:
The Making and Unmaking of Ignorance. Stanford,
CA: Stanford University Press.
Proctor RN and Schiebinger L (2008) Agnotology: The
Making and Unmaking of Ignorance. Stanford, CA:
Stanford University Press.
Puigde la Bellacasa M (2014) Encountering bioinfrastructure:
ecological struggles and the sciences of soil. Social
Epistemology 28(1): 26–40. DOI: 10.1080/02691728.
Rappert B (2015) Sensing absence: how to see what isn’t
there in the study of science and security. In: Rappert
B and Balmer B (eds), Absence in Science, Security
and Policy. Global Issues Series. London: Palgrave
Rayner S (2012) Uncomfortable knowledge: the social
construction of ignorance in science and environ-
mental policy discourses. Economy and Society 41(1):
107–125. DOI: 10.1080/03085147.2011.637335.
Robbins P, McSweeney K, Chhangani AK, et al (2009)
Conservation as it is: illicit resource use in a wildlife
reserve in India. Human Ecology 37(5): 559–575.
Roitman J (2006) The ethics of illegality in the chad basin.
In: Comaroff J and Comaroff JL (eds), Law and
Disorder in the Postcolony. Chicago: University of
Salisbury DS and Fagan C (2013) Coca and conservation:
cultivation, eradication, and trafﬁcking in the Amazon
borderlands. GeoJournal 78(1): 41–60. DOI: 10.
Samers M (2004) An emerging geopolitics of ’illegal’
immigration in the European Union. European
Journal of Migration and Law 6(1): 27–45. Brill
Nijhoff. DOI: 10.1163/1571816041518750.
Sanabria E (2016) Circulating ignorance: complexity and
agnogenesis in the obesity “epidemic”.Cultural An-
thropology 31(1): 131–158. DOI: 10.14506/ca31.1.07.
Sears RR and Pinedo-Vasquez M (2011) Forest policy
reform and the organization of logging in peruvian
amazonia. Development and Change 42(2): 609–631.
Sedgwick E (2008) Epistemology of the Closet.Berkeley,Los
Angeles and London: University of California Press.
Sesnie SE, Tellman B, Wrathall D, et al. (2017) A spatio-
temporal analysis of forest loss related to cocaine traf-
ﬁcking in Central America. Environmental Research
Letters 12(5): 054015. DOI: 10.1088/1748-9326/aa6fff.
Slack J and Campbell H (2016) On narco-coyotaje: illicit
regimes and their impacts on the US-Mexico Border.
Antipode 48(5): 1380–1399. DOI: 10.1111/anti.12242.
Smithson M (1989) Ignorance and Uncertainty: Emerging
Paradigms. New York: Cognitive ScienceSpringer-
Star SL and Bowker GC (2007) Enacting silence: residual
categories as a challenge for ethics, information systems,
and communication. Ethics and Information Technology
9(4): 273–280. DOI: 10.1007/s10676-007-9141-7.
Stel N (2016) The agnotology of eviction in south lebanon’s
palestinian gatherings: how institutional ambiguity and
deliberate ignorance shape sensitive spaces. Antipode
48(5): 1400–1419. DOI: 10.1111/anti.12252.
Taussig M. (1984) Culture of terror-space of death. roger
casement’sputumayo report and the explanation of
torture. Comparative Studies in Society and History
Tellman B, Magliocca NR, Turner BL, et al. (2020175) Un-
derstanding the role of illicit transactions in land-change
dynamics. Nature SustainabilityNature Publishing Group
3(3): 175–181. DOI: 10.1038/s41893-019-0457-1.
670 Progress in Human Geography 46(2)
Tsing AL (2003) Natural resources and capitalist frontiers.
Economic and Political Weekly 38(48): 5100–5106.
van Schendel W (2005) Geographies of knowing, geog-
raphies of ignorance: jumping scale in Southeast Asia.
Locating Southeast Asia 20: 275–307. Brill. DOI: 10.
Vandergeest P and Peluso NL (1995) Territorialization and
state power in Thailand. Theory and Society 24(3):
385–426. DOI: 10.1007/BF00993352.
Watts M. J. (2018) Frontiers: authority, precarity, and in-
surgency at the edge of the state. World Development
101: 477–488. DOI: 10.1016/j.worlddev.2017.03.024.
Woods KM and Naimark J (2020) Conservation as
counterinsurgency: a case of ceaseﬁre in a rebel forest
in southeast Myanmar. Political Geography 83:
102251. DOI: 10.1016/j.polgeo.2020.102251.
Wrathall DJ, Devine J, Aguilar-Gonz´
alez B, et al. (2020)
The impacts of cocaine-trafﬁcking on conservation
governance in Central America. Global Environ-
mental Change 63: 102098. DOI: 10.1016/j.gloenvcha.
Yarwood R (2007) The geographies of policing. Progress
in Human Geography 31(4): 447–465. DOI: 10.1177/
Ybarra M (2012) Taming the jungle, saving the Maya
Forest: sedimented counterinsurgency practices in
contemporary Guatemalan conservation. Journal of
Peasant Studies 39(2): 479–502. DOI: 10.1080/
Zolberg AR (1989) The next waves: migration theory for a
changing world. International Migration Review
23(3): 403–430. SAGE Publications Inc. DOI: 10.
Dr. Laura Dev is an Assistant Professor in Envi-
ronmental Sciences and Society at the University of
Wisconsin-Platteville. Her current research focuses
on Indigenous land politics, more-than-human re-
lations, and climate justice in the Amazon.
Dr. Karly Marie Miller is an independent scholar
with research interests in coastal political ecology,
focusing in particular on livelihoods and resource
management in the context of coastal conservation
and development projects. She received her Ph.D. in
Marine Science from the University of California,
Dr. Juliet N Lu is a Postdoctoral Research Fellow at
Cornell University in the Atkinson Center for Sus-
tainability and the Department of Global Develop-
ment. She is a political ecologist and China scholar
focused on Chinese overseas agribusiness invest-
ments, development aid, and sustainability
Dr. Lauren S Withey is an independent scholar
with interests in tropical forest conservation and
common property regimes. She received her Ph.D. in
Environmental Science, Policy, and Management
from the University of California, Berkeley.
Dr. Tracy Hruska’s past research has centered
in rangeland political ecology with a particular focus
on the nexus of rural economic development and
natural resource conservation. He received his
Ph.D. in Environmental Science, Policy, and
Management from the University of California,
Berkeley. He currently works in riparian
restoration on the lower Columbia River.
Dev et al. 671