Counting COVID-19 deaths: a moveable feast

You may recall the adage that there’s nothing certain in life except death and taxes (notice the “uncertainty” trope at work; see previous Research Hub entry 29 April 2023). Well, you can scrap the first of these. It seems that in the time of COVID-19 you are dead for some purposes but not for others, or dead for some jurisdictions but not for others (Harries 2020; Trabsky & Hempton 2020). Or, I’ve heard it said that in the time of COVID-19 death does not count at all. 

My interest in this topic was sparked by the marshalling of the phrase “underlying health conditions” in reports by public health officials in Australia in their daily briefings on COVID-19 case numbers and mortality figures during 2021 and early 2022. I couldn’t help thinking at the time that some people were identified in this way, as having “underlying health conditions”, to divert attention from the rising death toll. After all, if people who died had “underlying health conditions”, their deaths appeared to have less to do with COVID-19. An implied lower death count could then serve to reduce the seriousness of the pandemic and, subsequently, allow more “freedoms” from public health restrictions.

Research on the topic led to several discoveries. First, the phrase, “underlying health conditions” or “pre-existing health conditions”, was being used in similar ways elsewhere. In a Research Hub entry entitled “Becoming More Mortal” (22 Feb. 2022) I referred to the work of the Native American novelist, Louise Erdrich (2021), who was clearly as bemused as I was by the operation of the term.  Further research led me to discover, with my colleague Anne Wilson, that the phrase played a key role in one of the longest running policy disputes in American history, the design of a health care system that covered the health of the population (Bacchi and Wilson 2022). Alongside these topics, I began to ask how such a phrase came to be (possible) and how it related to conceptions of health and illness more generally – recognizable as WPR questions 2 and 3 (see Chart in Bacchi and Goodwin 2016: 20). This search took me to the topic of death certification, the topic I introduce in this entry and pursue in two subsequent entries. 

As you shall see, in this series of three entries, I make the case that there is a good deal at stake in the form and content of the current internationally standardized death certificate (WHO 2022). Specifically, applying the WPR questions, I develop the argument that the death certificate, in its design and use, functions as a governmental mechanism to undermine the social determinants of health (SDH). There is no suggestion here of deliberate intent.

 To support this claim, this entry reviews how COVID-19 deaths are recorded and counted. In the second entry, in a month’s time, I offer a brief genealogy of how the international death certificate came to be adopted worldwide, with a particular focus on contestation. In the third entry of the series, I show how death certificates can undermine SDH, and review the possibilities and limitations of attempts to redress this situation (e.g., through Z codes).  

My governmentality background leads me to approach this topic in a specific way. Death certificates are described as “inscription devices”, to use the language of Actor-Network theory. Like graphs and maps, they are technologies or governmental mechanisms that, because of their appealing perceptual simplicity, “make it increasingly difficult to disagree with the matter at hand” (Latour, 1987: 64–70 in de Boer et al. 2021: 400). “Government” in this usage has a broader meaning than conventional uses of the term that target specific state institutions. It involves the regulation and oversight of the behaviours (or conduct) of the population by a wide range of agencies, authorities, experts and professionals (Bacchi 2023).  

As an example of a governmental mechanism, Rowse (2009) shows how the current Australian census problematizes Indigenous peoples as part of a population binary, Indigenous and non-Indigenous. Such a problematization, says Rowse, provokes consideration of the sort of political claims such a statistical distinction facilitates or blocks. In the European context, Walters (2002) explores how the apparently innocuous policy affecting signposts at airports, in which passengers are directed to EU and non-EU queues, serves to designate and firm up both “Europe” as a “place” and the category of “the European”. Importantly, in these studies the interest is not in rhetorical framing of the “problem” but in the governmental practices that produce certain problematizations and their political implications. 

Following in this tradition of exploring the political implications of commonly endorsed governmental mechanisms – e.g., censuses, signposts – I want to open reflections on how current standardized death certificates play significant roles in how lives are lived and how social relations are organized. Following Koopman et al.’s (2018) work on birth certificates, I see the need to pay more attention to death certificates as governmental technologies that shape worlds. 

Applying WPR (“What’s the Problem Represented to be?”) to death certification

In the language of WPR, death certificates become “practical texts” “which are themselves objects of a ‘practice’ in that … they were intended to constitute the eventual framework of everyday conduct” (Foucault 1986: 12-13; Bacchi 2009: 34). As practical texts they are problematizations with all sorts of repercussions for government funding and the design of public health interventions. How, then, I want to ask, do current practices of death certification affect governing? What specific effects can be associated with the patterns in death certification practices? Are their grounds for rethinking the conventions surrounding death certification practices and death certificates? 

These questions can be addressed through applying a WPR analysis. I start from the death certificate as a practical text and consider how it produces “cause of death” (COD) as a particular sort of problem.

I ask: what is the “problem” of “cause of death” (COD) represented to be in current death certification practices? One terminological issue needs to be sorted out at this point. The focus in death certificates is on “cause of death” and, more specifically, on the “underlying cause of death” (UCOD). This term needs to be kept distinct from “underlying health conditions”, which are not considered to be the “underlying cause of death”. Indeed, “underlying health conditions” are set apart from the “underlying cause of death”, both literally and figuratively, as explained in due course. To avoid confusion, I refer to “underlying health conditions” as “pre-existing health conditions”, which is often used as a synonym. 

As “practical texts”, death certificates provide the starting point for analysis. Recall that WPR examines proposals (proposed solutions, recommendations, etc.) in texts because they indicate what is identified as needing to change and hence what is produced as “the problem” (Research Hub 30 Jan 2023).

At one level, you could argue that the simple creation of a death certificate produces the “problem” as lack of certification, opening up consideration of the manifold reasons put forward for the usefulness of such certification. Historically, the practice of death certification dates to the 12th century and is considered to be a legal obligation of the attending doctor prior to the disposal of the decedent’s remains (Swift and West 2002). With the rise of nation states, emphasis is placed on standardized classifications so that information can travel across boundaries/borders (Bowker 1996: 50).

Going further, it becomes possible to examine the specific form of the death certificate – how it is designed, which items are deemed to be relevant – to understand what is being problematized. Here it becomes relevant to examine the changes made to the standardized death certificate over time. Such changes reveal shifts in approaches to health and illness. 

The current standardized international death certificate has gone through several iterations (Erhardt 1956; Alharbi et al. 2021). It follows the model recommended by the World Health Organization (WHO) in 1927 (see Stirton and Heslop 2018), and finally accepted in 1940 (Armstrong 1986: 219). Called the Medical Certificate of Cause of Death (MCCD), it consists of two parts directly targeting “cause of death”. Part 1 offers spaces for a doctor or a coroner to fill in the immediate cause of death on the first line and the “chain of events” leading to this outcome on subsequent lines. On the last line in this sequence (of usually three items) appears the “underlying cause of death” (UCOD), defined by the WHO as “a) the disease or injury which initiated the train of morbid events leading directly to death, or b) the circumstances of the accident or violence which produced the fatal injury” (WHO, 2010: 31 in Stirton and Heslop 2018: 659). Part 2 of the form is used to list “other significant conditions, diseases or injuries that contributed to the death, but were not part of the direct sequence leading to death” (Stirton and Heslop 2018: 660). It is here, in Part 2, that our “underlying [or pre-existing] health conditions” are placed – a location clearly distinct from the “underlying cause of death” and the “chain of morbid events”. 

Question 2 in WPR targets deep-seated assumptions and presuppositions in the design of death certificates. This question opens a plethora of themes that require elaboration. Let me run through two key themes: first, the conception of disease and hence the conception of the body in the death certificate; and second, the place of causation in the design of the form.  

On the first theme, Jewson’s (2009) notion of “medical cosmology” proves useful. Described as a “parallel idea” to Foucault’s discursive formations, medical cosmologies “are conceptual structures which constitute the frame of reference within which all questions are posed and all answers are offered”. Jewson describes the cosmological system of Hospital Medicine in these terms:

“At the centre of the new medical problematic was the concept of disease. Interest in the unique qualities of the whole person evaporated to be replaced by studies of specific organic lesions and malfunctions. Diseases became a precise and objectively identifiable event occurring within the tissues, of which the patient might be unaware. The fundamental realities of pathological analysis shifted from the total body system to the specialized anatomical structures. The experiential manifestations of disease, which had previously been the very stuff of illness, now were demoted to the role of secondary signs. The patient’s interest in prognosis and therapy was eclipsed by the clinician’s concern with diagnosis and pathology. The special qualities of the individual case were swallowed up in vast statistical surveys.” (Jewson 2009: 628)

This analysis indicates that the targeting of specific “organ systems” in death certificates needs to be historicized (WPR Question 3) and questioned rather than simply taken for granted. 

Second, we need to reflect on the conception of causation at work in death certificates. The focus is on a single underlying cause that operates through a “chain of events”, with causal links all the way up the chain to the final, “immediate” cause of death. The “causal chain” was introduced to deal with the tendency among doctors to list more than one underlying cause of death. The “chain” allowed a more complex understanding of the death process, without removing the focus from the “underlying cause”, which was deemed to be useful for epidemiological purposes (see next Research Hub entry). 

I have written elsewhere about the contested space around conceptions of causality, specifically about how causality is treated in Critical Realism (Bacchi 2016: 6). In this tradition, Pawson et al. (2005) describe how, to make a “causal inference” between “two events (X and Y),” “one needs to understand the underlying mechanism (M) that connects them and the context (C) in which the relationship occurs” (p. S1:21- 22). Mechanisms, in Critical Realism, refer to hypotheses about individuals’ behaviours, illustrating a basic methodological individualism. 

By contrast, a Foucault-influenced poststructural analytic strategy challenges conventional views of causality, in which one thing (or a few things) causes another. Instead, Foucault effects “a sort of multiplication or pluralization of causes” (Foucault 1991: 76), a proliferation of “events” as the random results of “the interweaving of relations of power and domination” (Tamboukou 1999: 207). “Everything depends on everything else” (Veyne 1997: 170). This perspective dramatically opens the understanding of the death “event” (Ariès 1982: 587), with implications for policy design – as we shall see when I address the space or lack of space for SDH in death certificates.  

Question 3 in WPR asks how we have arrived at this point in our conceptualization of COD. I pursue this question in the next entry with a particular focus on contestation. As a genealogy the task is to indicate that the death certificate might have developed differently, “to show that things ‘weren’t as necessary as all that’” (Foucault 1991: 76). 

Question 4 in WPR asks what is silenced through death certification practices. This question usually attracts the most attention among researchers. It is the question that encourages us to think “outside the square”, to imagine things differently. There are overlaps between this question and Question 5 on effects (Bacchi and Goodwin 2016: 20-24). Specifically, the category of “discursive effects” in Question 5 encourages us to reflect on how the current frame of reference for COD – the focus on diseases and organic “malfunctions” (Jewson 2009: 628) – makes it difficult to raise other “matters of concern” (Latour 2004), e.g., social determinants of health (SDH). I pursue this theme through the next two Research Hub entries, highlighting how the current international death certificate produces little to no room for reflecting on the social determinants of health (SDH).

Question 6 in WPR creates space to examine closely the specific practices involved in the development and use of standardized death certificates. Step 7 meanwhile confronts researchers (in this instance, me) with the need to consider the assumptions that underpin their (my) proposals for revising death certificates. For example, as someone who has needed numerous biomedical interventions, I find myself shifting between a questioning of the biomedical paradigm and sheer wonder at its successes. How to manage this tension becomes an important theoretical consideration.   

A good deal of contemporary discussion about death certificates focuses on how they are filled in. That is, concern is raised about the poor training offered medical residents and how they frequently get the COD “wrong” (Morgan 2022; McGivern et al. 2017). Without diminishing the significance of this lapse in protocol, a WPR approach directs criticism at another level. Schultz (2014) puts it this way: more important than the errors in reporting is “what we are looking for” and, by implication, what we are NOT looking for – what we fail to recognize as significant. In this and the subsequent two entries, I develop the argument that standardized death certificates do NOT look for the social and environmental causes of death. SDH is missing from death certificates and missing from consideration. The implications (effects as in Question 5 of WPR) are considerable. Again, Schultz (2014) pointedly states that “we count what we care about and more disturbing we care about what we count”. It’s time therefore to return to COVID-19 to see what we count.

COVID-19 deaths: What do we count? 

Historically and today, as I explain in more detail in the next entry, epidemics are treated as a separate category of disease. Consistently, they are described as underlying causes of death, pure and simple. However, as the concern about pandemic infections declined over the last century more and more attention has been directed to the “sequence of events” or “chain of morbid events” leading to death – the goal here to better understand the complex interplay of diseases, the “multiple causes”, that lead to death. The Australian biometrician, Treloar (1956: 1378), called for a “deeper study of the endemiology of chronic disease” due to the “passing of the era of acute communicable disease” (see also Moriyama 2011). And then COVID-19 arrived! 

The WHO responded as expected, given the historical treatment of epidemics as exceptional diseases. It lay down the rule to list COVID-19 as the “underlying cause of death”: 

“A death due to COVID-19 may not be attributed to another disease (e.g., cancer) and should be counted independently of pre-existing conditions that are suspected of triggering a severe course of COVID-19.” (WHO 2020: 3; emphasis added) 

A “causal sequence leading to death” is still required: “For example, in cases when COVID-19 causes pneumonia and fatal respiratory distress, both pneumonia and respiratory distress should be included, along with COVID-19, in Part 1”. However, COVID-19 is listed last in the “chain of events” as the “underlying cause”. If the person who dies had “existing chronic conditions”, also described as “comorbidities” or “underlying health conditions”, these are to be reported in Part 2 of the MCCD (Medical Certificate of Cause of Death).

These guidelines are applied in many countries (Australian Bureau of Statistics (ABS) 2020; Veeranna and Rani 2020). In Australia the ABS stresses the need to specify the “causal pathway leading to death in Part I of the certificate”, noting that “all conditions and symptoms should be included”. In the United States initially COVID-19 was treated like pneumonia, the final endpoint, or immediate cause, in the train of events leading to death. From April 2020 the new rules required COVID-19 to be listed last as the underlying cause. Therefore, the train of events was reorganised: “any long-term conditions, no matter how serious, were then relegated to Part II of the Death Certificate as ‘contributing’ causes” (Kiang et al. 2020).

The listing of a “sequence of events” leading to death from COVID-19 and the inclusion of “underlying [pre-existing] health conditions” in Part 2 of the death certificate led to a debate about whether people died “with” COVID-19 or “from” (“of”) COVID-19. The Queensland Government (2022) uses this distinction – a distinction not found in the WHO Guidelines – in their information for health professionals about recording and reporting COVID-19 deaths:

         “Recording a death “from” COVID-19

If the underlying cause of the death is COVID-19, then “COVID-19” should be recorded in Part 1 of the Medical Certificate of Cause of Death and be included on the lowest line as the underlying cause with all antecedent conditions and symptoms (including duration) that led to the disease or condition resulting in death. Other significant considerations contributing to the death but not related to the diseases or conditions causing the death are recorded in Part 2

Recording a death “with” COVID-19

If a person had COVID-19 but the virus was not part of the chain of events leading to death, COVID-19 should be included in Part 2 “other significant conditions contributing to death”, with the main condition [sic] disease or condition in Part 1 followed by the underlying causes below.”

A recent report from the Australian Actuaries COVID-19 Mortality Working Group (Actuaries Digital 2023) noted that deaths previously referred to as “with COVID-19”, are now described as “COVID-19 related”, meaning that COVID-19 “contributed to the death”. 

This distinction between dying “with” COVID-19 and dying “from” COVID-19, which is difficult to make, can be used to reduce the case fatality rate (CFR) and, hence, the significance accorded the pandemic (Trabsky 2020; Amoretti and Lalumera 2020). It prompted some to make the case that the appearance of other “causes” on the death certificate means that you cannot attribute the death to COVID-19.  This reasoning explains Trump’s figure of 6% for the percentage of deaths due to COVID-19 (Aschwanden 2020). These were the 6% of doctors or coroners who listed COVID-19 as the underlying and only cause of death, not mentioning the “chain of events” and “contributing” conditions. There are reports of undercounting of COVID-19 cases in the ACT and NSW due to the failure to count those where COVID-19 was not listed as a “contributing factor or cause of death” (Ferguson 2022). 

In reaction to this confusion, there has been increasing reliance on the calculation of “excess deaths” in estimating the impact of COVID-19. These figures are estimated by comparing observed death versus expected mortality rates based on prior years. Examining the “excess mortality” in Australia – i.e., the deaths above those that would have been expected had there been no pandemic – the Actuaries’ Report noted that: there were over 20,000 more deaths in 2022 than expected; just over half of these (10,300) were “due to deaths from COVID-19″ (identified as the “underlying cause” on death certificates); another +2,900 were “COVID-19 related deaths”, while 7,000 made no mention of COVID-19. The Report noted that COVID-19 was a likely catalyst in COVID-19 related deaths, and that it is difficult to know “how much ‘blurring’ there may be between deaths from COVID-19 and COVID-19 related deaths” (emphases in original). 

The Actuaries’ Report links the 7,000 excess deaths that did not mention COVID-19, in part, to an association between COVID-19 and subsequent higher mortality risk from heart disease: “certifying doctors would generally not identify a causative link several months after recovery from COVID-19”. Some of these excess deaths are also attributed to pandemic-related delays in emergency care and delays in routine care (Actuaries’ Institute 2023). These deaths are not considered to be due to COVID-19 since they do not involve medical mechanisms. 


I suggest that it is time to shift the focus of analysis from what we count in terms of COD to what we don’t count, putting into question the central parameters of the death certificate. Building on Schultz (2014), if we care about what we count, then it’s time to start counting more things as causing death – e.g., pandemic-related delays in emergency care and delays in routine care. 

This proposition operates at several levels, as will become clear in the next two entries. On one level I raise questions about the focus in death certificates on “disease entities” and what this focus precludes from analysis (i.e., the social determinants of health). At another, and perhaps more practical level, I consider whether it is feasible to bring the social determinants of health into death certification practices. I ask: do “multiple causes” approaches offer a possible way forward? And what about Z codes?

Before I broach these questions, I trace, in the next entry, some of the strands of influence that have brought us to the current situation in delineating causes of death. I identify a biology/environment tension in approaches to population health and suggest the difficulties the environmental option continues to face. 


Actuaries Institute 2023. COVID-19 Mortality Working Group: Confirmation of 20.000 excess deaths for 2022 in AustraliaActuaries Digital, 6 April.

Alharbi, M. A., Isouard, G. & Tolchard, B. 2021. Historical development of the statistical classification of causes of death and disease. Cogent Medicine, 8(1), 1893422. DOI: 10.1080/2331205X.2021.1893422

Amoretti, M. C. and Lalumera, E. 2020. COVID-19 as the underlying cause of death: disentangling facts and values. Hist Philos Life Sci, 43(4).

Ariès, P. 1982. The Hour of Our Death. NY: Vintage Books. 

Armstrong, D. 1986. The invention of infant mortality. Sociology of Health & Illness, 8(3): 211-232.

Aschwanden, C. 2020. Debunking the False Claim that COVID Death Counts Are Inflated. Scientific American, 20 October.  

Australian Bureau of Statistics (ABS) 2020. Guidance for Certifying Deaths due to COVID-19. Available at:

Bacchi, C. 2009. Analysing Policy: What’s the Problem Represented to be? Frenchs Forest: Pearson Education.

Bacchi, C. 2016. Problematizations in Health Policy: Questioning How “Problems” Are Constituted in Policies. Sage Open, April-June, 1-16. DOI: 10.1177/2158244016653986

Bacchi, C. 2023. Governmentalizing “policy studies”. In W. Walters and M. Tazzioli (Eds) A Handbook of Governmentality, pp. 54-72. Edward Edgar.    

Bacchi, C. and Goodwin, S. 2016. Poststructural Policy Analysis: A Guide to Practice. NY: Palgrave Macmillan.

Bacchi, C. and Wilson, A. 2022. Governing through “underlying (pre-existing) health conditions”: “chronic illness”, “race-ism” and COVID-19. Paper presented at the International Symposium on Critical Policy Studies, 17-19 August 2022. Paper available on request (

Bowker, G. C. 1996. The History of Information Infrastructures: The Case of the International Classification of Diseases. Information Processing & Management, 32(1): 49-61. 

De Boer, B., te Molder, H. and Verbeek, P-P. 2021. Understanding science-in-the-making by letting scientific instruments speak: From semiotics to postphenomenology. Social Studies of Science, 51(3): 392-413. 

Erdrich, L. 2021. The Sentence. Harper.

Erhardt, C.L. 1958. What Is “the Cause of Death”? Journal of the American Medical Association, 168, 161-168.

Ferguson, K. 2022, NSW’s COVID-19 death toll rises by 331 after medical cross-checks, 1 April. Available at:

Foucault, M. 1986. The Use of Pleasure. The History of Sexuality, Volume 2. Trans. R. Hurley. London: Viking Press.

Foucault, M. 1991. Questions of method. In G. Burchell, C. Gordon, P. Miller (Eds.), The Foucault effect: Studies in governmentality. Chicago: University of Chicago Press.  

Harries, D. 2020. How do different countries calculate their COVID-19 death rates? CGTN, 18 April. Available at:–PM3ur6PEtO/index.html.

Jewson, N. D. 2009. The disappearance of the sick-man from medical cosmology, 1770-1870. International Journal of Epidemiology, 38: 622-633.

Kiang, M. V., Irizarry, R. A., Buckee, C. O., and Balsari, S. 2020. Every Body Counts: Measuring Mortality From the COVID-19 Pandemic. Annals of Internal Medicine, 173, 1004–1007.

Koopman, C., Sheehey, B., Jones, P., Smithers, L., Hamid, S., Pickard, C., & Critical Genealogies Collaboratory. 2018. Standard forms of power: Biopower and sovereign power in the technology of the US birth certificate, 1903-1935. Constellations,1-16. Online ahead of print. doi:10.1111/1467-8675.12372

Latour, B. 1987. Science in Action: How to Follow Scientists and Engineers Through Society. Cambridge, MA: Harvard University Press.

Latour, B. 2004. Why Has Critique Run out of Steam? From Matters of Fact to Matters of Concern. Critical Inquiry, 30(2): 225-248.

McGivern, L, Shulman, L., Carney, J. K., Shapiro, S. and Bundock, E. 2017. Death Certification Errors and the Effect on Mortality Statistics. Public Health Reports, 132(6): 669-675.

Morgan, A., Andrew, T., Guerra, S., Luna, V., Davies, L. and Rees, J. 2022. Provider reported challenges with completing death certificates: A focus group study demonstrating potential sources of error. PLOS ONE |

Moriyama, I. M. and Rosenberg, H. M. 2011.  History of the Statistical Classification of Diseases and Causes of Death. Centers for Disease Control and Prevention. National Center for Health Statistics. 

Pawson, R., Greenhalgh, T., Harvey, G., & Walshe, K. 2005. Realist review—A new method of systematic review designed for complex policy interventions. Journal of Health Services Research & Policy10(Suppl.1), 21-S1.34.

Queensland Health 2022. Information for health professionals about recording and reporting COVID-19 deaths. Queensland Government, 27 Sept.

Rowse, T. 2009. The ontological politics of “closing the gaps”. Journal of Cultural Economy, 2 (1&2): 33-48. 

Schultz, K. 2014. Final forms: What death certificates can tell us, and what they can’t. New Yorker, 7 April.

Stirton, F. D. and Heslop, P. 2018. Medical Certificates of Cause of Death for people with intellectual disabilities: A systematic literature review. Journal of Applied Research in Intellectual Disabilities, 31: 659-668. 

Swift, B. and West, K. 2002. Death certification: An audit of practice entering the 21st century. Journal of Clinical Pathology, 55: 275-279.

Tamboukou, M. 1999. Writing genealogies: An exploration of Foucault’s strategies for doing research. Discourse: Studies in the Cultural Politics of Education, 20 (2), 201–217.

Trabsky, M. and Hempton, C. 2020. “Died from” or “died with” COVID-19? We need a transparent approach to counting coronavirus deaths. The Conversation, 9 September.

Treloar, A. E. 1956. The Enigma of Cause of Death. Journal of the American Medical Association (J.A.M.A.), p.8.

Veeranna, C. H. and Rani, S. 2020. Cause of Death Certification in COVID-19 Deaths. Indian Journal of Critical Care Medicine, 24(9): 863-867.

Veyne, P. 1997. Foucault revolutionizes history. In A.I. Davidson (Ed.), Foucault and his Interlocutors, Porter, C. (trans.), Chicago: University of Chicago Press.

Walters, W. 2002. The power of inscription: Beyond social construction and deconstruction in European integration studies. Millennium: Journal of International Studies, 31(3): 83-108.  

World Health Organisation 2010. International Statistical Classification of Diseases and Related Health Problems 10th Revision Volume 2 Instruction manual 2010 Edition. Retrieved from classifications/icd/ICD10Volume2_en_2010.pdf

World Health Organization (WHO) 2020. International guidelines for certification and classification (coding) of COVID19 as cause of death  

World Health Organization 2022. ICD-11  Mortality tooling. Available at:

Applying WPR to concepts: questioning “risk”, “uncertainty” and “crisis”

It seems to be a truism that, in the throes and anguish of a pandemic, we are living in a time of “crisis”. The term certainly appears often enough in the titles of published articles and could be designated a trope of our times. Applying WPR thinking, I suggest the need to stand back from the automatic adoption of such tropes to examine them as governing or governmental mechanisms. To develop this argument, I consider the primacy accorded “crisis” in Koselleck’s Begriffsgeschichte, the links between “crisis” and “risk” as governing technologies, and the invocation of “uncertainty” as a new model for doing politics. 

Categorizing approaches to “crisis”

To begin, it is useful to offer examples of the most common ways in which “crisis” operates as a concept in current academic debates. I identify three conceptualizations: first, the scientific conception of “crisis” as “truth”; second, the identification of “crisis talk” as rhetorical manipulation; and third, the production of “crisis” as social construction. The poststructuralist option I proceed to develop evades these three options and engages “crises” as governmental problematizations. First, let us look briefly at the three more common uses of the term. 

  •  The psychologist Dev Roychowdhury (2020) makes the scientific case for “a comprehensive understanding or consensus on a definition of crisis”. He argues that such an understanding “is not only a part of sound academic approach, but also paves the ways for scientific investigation to be undertaken with the aim of explaining, predicting, and managing crisis successfully”. A firm definition of crisis, in this view, is necessary for science to be accredited as a guide to managing present and future challenges. It follows that to achieve this goal requires “access to up-to-date evidence-based guidance” (Odium et al. 2021: Abstract).
  •  Molla and Cuthbert (2022) direct attention to “crisis exploitation” (Boin et al. 2008) as a form of rhetorical manipulation. They describe how the Australian Government “framed the COVID-19 pandemic and the associated fear of an economic downturn as a crisis to sell old policy packages and imaginaries” (Molla and Cuthbert 2022: Conclusions). Here we have a clear example of the argument that concepts can be marshalled for political ends. 
  •  Aligned to this view a crisis can be treated as if it were socially constructed through narrative (Hay 1996) or “discourse” (Walby 2022: 499). For example, Agamben (2020) is concerned that “the crisis has been exaggerated, if not invented, for the purpose of legitimating the concentration of power in the hands of the executive branch of government” (in Walby 2022: 507). 

Walby (2022: 501) usefully points out that “crisis” is commonly treated as “real” – i.e., the scientific option – in fields of risk (Beck 1992, 2009; discussed below) whereas the notion that a “crisis” is socially constructed is common in the field of security, “especially when the issue is centred on a threat rather than physical harm”. She interprets this usage to mean that “a crisis is both real and its meaning is contested” (Walby 2022: 501). Approached through WPR thinking, a crisis is neither real nor “unreal”, and acknowledging disputes about its meaning does not advance us very far theoretically (see Hall 1998: 80). Rather, attention turns to the ways in which the concept “crisis” influences “the evolution of ongoing practices” (Tanesini 1994: 207).

Conceptual traditions and “crisis” 

Where do the conceptual traditions introduced in the first two entries in this series (27 Feb 2023; 30 March 2023) fit within this schema of scientific, rhetorical or social constructionist approaches to “crisis”? As mentioned in the previous entry and developed below there is some blurring of the boundaries between the Cambridge School and Begriffsgeschichte on the treatment of “crisis” as “speech act”. At the same time, the attention Koselleck, and others associated with Begriffsgeschichte (Jordheim and Wigen 2018), direct towards what they call the “conceptual work” performed by concepts suggests some possible overlap with a poststructural performative perspective. The task, as I see it, is to clarify the relationship between “conceptual work”, as described in Begriffsgeschichte, and WPR, which I proceed to do below. To forecast the argument, Begriffsgeschichte focuses primarily on political actors who marshal concepts, such as “crisis”, for political purposes, whereas WPR examines the operation of such concepts within governmental directives, read broadly. 

The term “crisis” is central to Koselleck’s work, introduced in the first entry in this series on WPR and concepts (27 Feb. 2023).  Koselleck’s doctoral thesis from 1959, later published under the title Kritik und Krise (Criticism and Crisis), offers a critique of Enlightenment philosophers (Voltaire, Diderot, Rousseau, Schiller, and Kant) for evoking “crisis” to avoid “direct political confrontation with the absolutist state”. Instead, in his view, they submitted the absolutist state to a form of moral criticism “deemed to bring about its collapse”. In this way, argues Koselleck, “to talk of crisis is to bring crisis about” (in Jordheim and Wigen 2018: 438).

Jordheim and Wigen (2018: Abstract) build on Koselleck’s ideas to develop an argument about the ways in which concepts – with their targets being “progress” and “crisis” – are used to order events, objects and polities. They argue that, since the late 18th century, “the key concept for structuring the relationship between past, present, and future in Western culture has been progress” (Jordheim and Wigen 2018: 425). Their claim is that “the concept of crisis is about to replace the concept of progress as the main tool of historicization in the Western world and beyond”. In their view, what happens with these concepts is vitally important for the future world order: “which one [i.e., progress or crisis] is allowed to dominate world politics is highly decisive for what we are able to experience, to plan and to do”. 

Jordheim and Wigen (2018: 438) draw a link between their argument on the important role played by concepts and the Cambridge school’s use of “speech act” theory (see discussion in Research Hub 27 Feb 2023) They state that “labelling something a crisis can be thought of as a speech act that shapes political contestation over a particular issue or field”. This outcome, they argue, is due to the work that concepts do. To illustrate this point, using the example of the concept “security”, they describe how “managing to label a particular issue, topic or field a matter of security excludes a great many actors who would otherwise be party to discussion from legitimately debating it” (Jordheim and Wigen 2018: 438). They also describe how crisis functions as a “collective singular” (see Research Hub 27 Feb. 2023), “with a capital C if you like” (Jordheim and Wigen 2018: 431).  

As with the Cambridge school the focus in this account is on social actors and their deployment of concepts, though Jordheim and Wigen insist that they are interested in more than “rhetorical skills” (Jordheim and Wigen 2018: 425). The starting questions for their analysis, however, target language users and their motivations: “Who uses the word? In what way? What does it mean? What does the user want to achieve?” (Jordheim 2021). And they explicitly identify their concerns as the ways in which political leaders use concepts: “To organise and take control over time, state leaders, politicians and intellectuals of statecraft use concepts as tools of synchronisation” (Jordheim and Wigen 2018: 439). 

This focus on language use can also be seen in the ways in which the authors refer to “discourse/s”. For example, they describe “calling something security” and “calling a particular issue or set of events a crisis” as two related “discursive moves” (Jordheim and Wigen 2018: 439; emphasis in original). Such a reference indicates that “discourse” here is conceptualized as language use rather than as knowledges, illustrating the important distinction between these positions introduced in the previous Research Hub entry.  Concepts in Begriffsgeschichte are up for grabs and serve the specific political purposes of “discourse participants”, akin to the position developed by Molla and Cuthbert (2022; see above).  

What does WPR do differently? The three Research Hub entries (27 Feb 2023; 30 March 2023; and today’s) on this difficult topic have a single goal – to suggest that WPR opens up new terrain for exploration when dealing with “concepts”. Approaching concepts as governmental problematizations leads to different foci of attention and to different questions. Rather than asking about who is using a concept or what they want to do with it (see above), the focus becomes the complex of forces and practices that go into a concept’s making, and the effects that accompany its production and use. Clearly, this stance puts in question any attempt to define “crises” as “real” and measurable (the scientific option). It also moves the analysis past local actors and their attempts to deploy “crisis talk” for political purposes (the rhetorical manipulation position). Finally, it does not imply that “crises” are “unreal” and socially constructed. Rather, it asks how the concept functions to shape realities and possible futures.

Here there is some overlap with the notion of “conceptual work” developed by Jordheim and Widen (2018). They offer useful insights into the ways in which the concepts of “progress” and “crisis” can be seen to shape political possibilities. However, where they target the people who deploy concepts, WPR interrogates how concepts operate to shape realities within governmental directives (guides to conduct) – with governmental understood to encompass wide-ranging societal administration (Dean 1999). To put it starkly, the concepts that are the focus of analysis in WPR are to be found in governmental accounts and records, not in people’s heads. Hence, I argue, interrogating them provides insights into how governing takes place that are unavailable if the target is restricted to politicians’ and intellectuals’ “statecraft” (see above). 

The link drawn by Jordheim and Wigen between Koselleck’s history of concepts and Austin’s speech act theory, associated with the Cambridge school, raises the whole issue of “illocutionary forces”, and hence of performativity theory (Research Hub, 29 Sept. 2022, 26 Oct. 2022). In the Keynote Address at Karlstad ( KEYNOTE ADDRESS – CAROL BACCHI – 18 August 2022) I suggest that proposals, recommended as starting points for WPR analyses, are “in a sense analogous to Austin’s (1962) illocutionary performatives that do not describe ‘reality’ but that (help to) make worlds”. I need to stress the word analogous in this argument given the ways in which, in both Koselleck and the Cambridge group, speech act theory is linked to “human agency as the prime mover of history” (Van Gelderen 1998: 231). Koselleck confirms this interpretation of his work with his description of concepts as “linguistic performances” (Koselleck 2004: 232). By contrast, the focus of analysis in WPR is not “linguistic performances”; rather, its target consists of governmental directives, read broadly. This position can be illustrated by reflecting briefly on how WPR approaches the topics of “crisis”, “risk” and “uncertainty”. 

Standing back from “crisis”, “risk” and “uncertainty”

Above I noted that “crisis” is commonly treated as “real” (scientific) in fields of risk. This alignment can be explained through the ways in which “risk” features in so many aspects of our lives – consider health, insurance, “risk management” technologies, and so on. Beck (1992, 2009) is best known for theorizing the “risk society”. As Rose (2000: 246) describes, in Beck there is an acceptance that risk reflects “changes in the contemporary existential conditions of humans and their world”. In other words, “risk” is treated as something we have to deal with and live with – “risks” are real. 

By contrast WPR and other governmentality analyses approach “risk” as a governmental strategy, with governmental read broadly, rather than as an inescapable part of contemporary lives. Rose (2000: 246; emphasis in original) notes that “genealogical studies have analysed risk as a particular style of thinking born during the nineteenth century”.  As he explains, “risk thinking brought the future into the present and made it calculable”. He notes a shift from an earlier period of the collectivization of risk to a contemporary individualization of risk.

Also, within a governmentality framework, Nettleton’s (1997) work on risk categories in health highlights how an individualizing logic is at work in health policy, with important subjectification effects. Says Nettleton, we are produced as “risky selves”, responsible for continuous scrutiny of diet and lifestyle to avoid threats to health (Nettleton 1997). 

This example illustrates how approaching the concept of “risk” as a problematization creates the space to reflect on the ways in which we are governed through “risk” categories and “risk” technologies. Importantly, there is no presumption that people are passive in response to such directives. As Nettleton (1997: 217) asserts, “discourses on health and health policy presuppose a self that is able to react to and challenge ‘expert’ knowledge”.  

Developing the governmentality position, Dean (1999: 177) states that “There is no such thing as risk in reality”:

“Risk is a way – or rather, a set of different ways – of ordering reality, or rendering it into a calculable form. It is a way of representing events in a certain form so they might be made governable in particular ways, with particular techniques and for particular goals”.

The objective for researchers, therefore, is to ascertain the interconnected sets of practices involved in the production of “risk” thinking and “risk technologies”, including practices of emergence, insertion/institutionalization and functioning (Foucault 1972: 163; 1991: 65). Dean (1999: 178) elaborates this research agenda clearly: 

“What is important about risk is not risk itself. Rather it is: the forms of knowledge that make it thinkable, such as statistics, sociology, epidemiology, management and accounting; the techniques that discover it, from the calculus of probabilities to the interview; the technologies that seek to govern it, including risk screening, case management, social insurance and situational crime prevention; and the political rationalities and programmes that deploy it, from those that dreamt of a welfare state to those that imagine an advanced liberal society of prudential individuals and communities”.

WPR suggests treating “crisis” in precisely this way. In research, the task is not to identify “crises” but to examine how the concept “crisis” operates to shape political possibilities and political outcomes. For example, Larsson (2020: Abstract) examines the “preemptive logic of contemporary security and crisis management” to show how “civil and war preparedness are merged into an ever-present dimension of everyday existence”. 

In an earlier Research Hub entry (31 March 2021) I considered how the concept of “uncertainty” can be examined as a problematization and a governmental technology. “Uncertainty” is, of course, a close cousin of “crisis” and “risk”. We are constantly being told that we live in uncertain times – though it is uncertain (forgive me!) how that observation is to assist us. 

Standing back from “uncertainty” as a presumption about our existential condition allows us to consider its implications for how society is ordered and governed. Pre-COVID-19 Pellizzoni (2011) made the case that uncertainty has become a way of governing, a technology of governing: “Uncertainty, thus, is seen no more as a circumscribed situation on which to build a few strategic decisions, but as an empowering everyday condition”. Describing the orientation as neoliberal, he notes: “proper calculations of risk are seen as an exception, while reasoned bets over unpredictable futures are regarded as the rule”. 

These “reasoned bets” produce a style of governing based on experimentation. As I described in the earlier entry (Research Hub 31 March 2021), experimenting has become increasingly normalized as a mode of governing. That is, it has been generated as a legitimate way to govern, as a kind of “truth” of governing.

This style of governing has important effects. Schroth (2016) makes the point that experiments reduce the “enigmatic world” to what are deemed to be manageable proportions. Hence, we need to consider what gets left out – what is silenced. In relation to COVID-19 the focus on “uncertainty”, “crisis” and “risk” produces a tendency to concentrate on what Aly (2020) calls “the symptoms” of a crisis. There is no space in this reactive approach to consider how we have got here – how pandemics, for example, are an increasing likelihood due to a range of practices such as deforestation (Zimmer 2011). At the same time, the invoking of “uncertainty” works paradoxically to “reinforce the authority of expertise” (Demeritt 2001: 327), those who will guide us through “uncertain” times. 


In this and the preceding two entries, I have explored the proposition that we are governed through the ways in which concepts problematize issues and that, hence, applying WPR to concepts produces new forms of questioning and analysis. I hope the examples of “crisis”, “risk” and “uncertainty” successfully illustrate the kinds of questions that arise in WPR thinking, and how these are vitally important questions for considering how we govern and how governing takes place.

Clearly, this position on concepts has repercussions for our own research. It imposes an obligation to consider the concepts we use, where they come from and how they operate politically. This self-interrogation can, doubtless, be uncomfortable (Foucault 2000). Some might also query the usefulness of such self-problematization. The claim advanced on behalf of this poststructural theorizing is that there is a need to reflect critically on the frameworks of meaning we may inadvertently adopt. Such interrogation creates the space to explore alternative problematizations, the space to think differently. We can only be the beneficiaries of such endeavours. 


Agamben, G. 2020. The invention of an epidemic, coronavirus-and-philosophers/ (originally published in Italian on Quodlibet,www.

Aly, W. 2020. Get to the root of mess. The Melbourne Age, Saturday 26 Dec., p. 48.

Austin, J. L. 1962. How to do things with words. Massachusetts: Harvard University Press.

Beck, U. 1992. Risk Society: Towards a New Modernity, London: Sage.

Beck, U. 2009. World at Risk, Cambridge: Polity.

Boin, A., ‘t Hart, P., & McConnell, A. 2008. Conclusions: The politics of crisis exploitation. In A. Boin, A. McConnell & P. ‘t Hart. (Eds.), Governing after crisis (pp.285 – 216). Cambridge University Press.

Dean, M. 1999. Governmentality: Power and rule in modern society. London: Sage.  

Demeritt, D. 2001. The Construction of Global Warming and the Politics of Science. Annals of the Association of American Geographers, 91(2): 307-337.

Foucault, M. 1972. The Archaeology of Knowledge, translated by A. Sheridan Smith. New York: Pantheon Books. 

Foucault, M. 1991. Politics and the study of discourse. In G. Burchell, C. Gordon & P. Miller (Eds), The Foucault Effect: Studies in Governmentality. Chicago: University of Chicago Press.

Foucault M. 2000a. For an ethics of discomfort. In J. D. Faubion (ed.) Power: Essential Works of Foucault, 1954– 1984, Volume III. New York, NY: The New Press. pp. 443–448.

Hall, T. 1998. Conceptual History and the History of Political Thought. In I. Hampsher-Monk, K. Tilmans and F. van Vree (Eds) History of Concepts: Comparative Perspectives. Amsterdam: Amsterdam University Press. pp. 75-86.

Hay, C. 1996> Narrating crisis: the discursive construction of the ‘Winter of Discontent’. Sociology, 30(2): 253–77. doi: 10.1177/0038038596030002004 

Jordheim, H. and Wigen, E. 2018. Conceptual Synchronisation: From Progress to Crisis.  Millennium: Journal of International Studies, 46(3): 421-439.

Jordheim, H. 2021. Conceptual History. In Bloomsbury History: Theory and Method.DOI: 10.5040/9781350970878.066

Koselleck, R. 2004. Futures Past: On the Semantics of Historical Time. Trans. K. Tribe. Columbia University Press. 

Larsson, O. L. 2020. The connections between crisis and war preparedness in Sweden. Security Dialogue, 1-19.

Molla, T. and Cuthbert, D. 2022. Crisis and policy imaginaries: higher education reform during a pandemicHigher Education, 

Nettleton, S. 1997. Governing the risky self: How to become healthy, wealthy and wise’. In A. Petersen and R. Bunton (Eds) Foucault, Health and Medicine. London: Routledge.  

Odium, A. et al. 2021.  Use of COVID-19 evidence in humanitarian settings: the need for dynamic guidance adapted to changing humanitarian crisis contexts. Conflict and Health, 15:83

Pellizzoni, L. 2011. Governing through disorderz: neoliberal environmental governance and social theory. Global Environmental Change, 21(3), 795-803.

Rose, N. 2000. Powers of Freedom: Reframing political thought. Cambridge, UK: Cambridge University Press. 

Roychowdhury, D. 2020. 2019 Novel Coronavirus Disease, Crisis, and Isolation. Frontiers in Psychology, doi: 10.3389/fpsyg.2020.01958 

Schroth, F. 2016.The Politics of Governance Experiments: Constructing the Clean Development Mechanism. PhD thesis, Berlin Institute of Technology. 

Tanesini, A. 1994. Whose language? In K. Lennon, & M. Whitford (Eds.), Knowing the difference: Feminist perspectives in epistemology. New York: Routledge. 

Van Gelderen, M. 1998. Between Cambridge and Heidelberg. Concepts, Languages and Images in Intellectual History.

In I. Hampsher-Monk, K. Tilmans and F. van Vree (Eds) History of Concepts: Comparative Perspectives. Amsterdam: Amsterdam University Press. pp. 227-238.

Walby, S. 2022. Crisis and society: developing the theory of crisis in the context of COVID-19. Global Discourse, 12(3-4): 498-516. Special Issue: Critical Explorations of Crisis: Politics, Precariousness, and Potentialities.  Zimmer, C. 2011. A Planet of Viruses. University of Chicago Press

Applying WPR to concepts: “analysis of discourses”, not “discourse analysis”

Discourse traditions

In the previous entry I developed the argument that concepts ought to be treated as governmental mechanisms, centrally involved in how governing takes place. This description follows Dean’s (1999: 44-45) proposition that political vocabulary can be usefully viewed as a governmental technology due to its role in making “politics thinkable”. By approaching concepts in this way, they are “denaturalised, made specific and their governmental implications revealed” (Larner 2008: 23).

How then are concepts related to “discourse” and “discourses”? 

This topic became relevant while I was researching the German tradition of Begriffsgeschicht (History of Concepts), introduced in the last entry (27 Feb 2023)An important article by Krzyzanowski (2010) suggests the theoretical benefits of bringing together Critical Discourse Analysis (CDA) – or more precisely Discourse-Historical Analysis (DHA) – and Koselleck’s  Begriffsgeschichte. Krzyzanowski (2010: 128) makes the case that “discourses” in DHA equate to “concepts” in Begriffsgeschicht.

This argument relies on a particular understanding of “discourse”, an issue that comes up repeatedly when considering the relationship between WPR and CDA/DHA. Many researchers have adopted the practice of developing a dual theoretical stance, combining WPR with CDA. Some time ago I discussed why I have concerns regarding this blending of theoretical approaches (Research Hub, 14 May 2018). As the practice continues, however, I feel the need to revisit the topic. The issue is also clearly connected to how “concepts” are conceptualized, our theme in this series of entries. 

To state the case clearly, I argue that WPR is best described as “an analysis of discourses” rather than as “discourse analysis” (Bacchi 2005). “Discourse analysis” involves the study of language use. By contrast, “an analysis of discourses” involves the study of discourses, understood as “knowledges”. I developed this distinction in a 2005 article titled provocatively, “Discourse, Discourse Everywhere: Subject ‘‘Agency’’ in Feminist Discourse Methodology”. I can remember at the time feeling overwhelmed by the ubiquity of the term “discourse” in the academic literature with which I engaged. It became clear to me that researchers were not always referring to the same thing when they spoke about “discourse”. 

I distinguished two central analytic traditions in discourse theory, a social psychological focus on patterns of speech (discourse analysis), and a political theoretical focus on the ways in which issues acquire a particular meaning within a specific social setting (analysis of discourses) (Potter & Wetherell 1990; Burr 1995:164). As mentioned above, in the first tradition the term discourse means something very close to language. There is a focus on the “linguistic and rhetorical devices” used in the construction of a text (Burr 1995:184). In the second tradition the goal is to identify, within a text, “institutionally supported and culturally influenced interpretive and conceptual schemas (discourses) that produce particular understandings of issues and events” (Bacchi 2005: 199).

In the 2005 article I note that this distinction should not be drawn too sharply. Clearly, WPR is interested in language use; however, attention to the operation of binaries and concepts/categories is a supplementary exercise, set in train through a focus on problematizations (Bacchi 2009). At the same time discourse analysts often pay heed to the interpretive and conceptual schemas that form the primary focus among those interested in the analysis of discourses. 

Nor am I suggesting that there is one correct definition of “discourse”. As discussed in the last entry (27 Feb. 2023), poststructuralists do not defend one meaning of a concept over another but consider how different meanings rely on contrasting presuppositions and produce diverse effects. For a poststructuralist, it is inconsistent, therefore, to search for a “correct” definition of discourse because the whole idea of discourse is that definitions play an important part in delineating knowledge (Bacchi 2000). Hence, definitions require scrutiny, not replication. 

What are these knowledges/discourses that are scrutinized in a WPR analysis? In a Foucault-influenced poststructural analysis, discourses/knowledges comprise both general background knowledge, apparent in epistemological and ontological assumptions, and forms of relatively bounded social knowledges, such as disciplines (Bacchi and Goodwin 2016: 21). Question 2 targets the epistemological and ontological assumptions that underpin specific problem representations. The concern with disciplinary knowledges appears in a particular emphasis on “expert” discourses. For example, the common reform proposal that women need training courses to gain access to higher paying and higher status occupations relies on a particular understanding of how “people” develop “skills” – a behavioural focus in other words – and an ontological conception of “people” as having “skills” ( KEYNOTE ADDRESS – CAROL BACCHI – 18 August 2022). In WPR, these assumptions, or presuppositions, underpin the knowledges identified as necessary to make the proposal to train women intelligible – e.g., psychological theories of development. These become the discourses/knowledges that require critical interrogation.

This distinction between “an analysis of discourses” and “discourse analysis” can be identified by examining how researchers in specific traditions use the term “discourse”.  For example, those who are interested in discourse as language use talk about “discursive practices” in terms of the ways in which people converse or argue. By contrast, in WPR, “discursive practices” refers to the ways in which discourses – or disciplinary knowledges – operate (or practice). Illustrating the latter stance, Foucault describes how, “in the nineteenth century, psychiatric discourse is characterized not by privileged objects, but by the way it forms objects that are in fact highly dispersed” (Foucault 1972: 44).  Therefore, when Foucault refers to “the discursive practice of psychiatry” (Foucault 1972: 75), he is not referring to how psychiatry is practised “discursively”, through writing, speaking or producing texts within psychiatry. His interest is not language use; rather, he is describing the practices of psychiatry, the operation of the sets of relations characteristic of psychiatry as an accredited form of knowledge (Bacchi and Bonham 2014: 182).

Other usages also indicate the two “discourse” traditions I have described here. When authors refer to discourses on or about an issue, e.g., climate change (Reisigl and Wodak 2017: 87, 119), this usage indicates a language perspective. The same is the case with references to “discourse participants” (Reisigl and Wodak 2017: 87), a phrase that points the way to a clearer delineation of the competing perspectives under examination here, as elaborated shortly. I need to repeat that these comments on the uses of “discourse” do not in any way suggest that one usage is correct and the other in error. Consistent with the argument developed in the previous entry (27 Feb. 2023) the task is to approach the contrasting uses of “discourse” in terms of how they offer contending proposals about how we ought to proceed from here (Tanesini 1994: 207).  

Krzyzanowski (2010: Introduction) indicates his positioning in this field through his recognition of and reliance on Ruth Wodak, commonly identified as a leading theorist in Discourse-Historical Analysis (DHA). Wodak (2015) and Reisigl and Wodak (2017) clearly lay out the theoretical parameters of DHA. Wodak (2015: 3) identifies the “ideas of Habermas” as “of fundamental importance”.  The goal of DHA is to improve “public discourse” through “a theory of rational argumentation” among “discourse participants” (Wodak 2015: 3). The goal of critique is to point to “intended biases in representations (especially media coverage) and to contradictory and manipulative relationships between discourses and power structures” (Reisigl and Wodak 2017: 119).

WPR sits in contrast to this position. It is not a form of discourse analysis in the tradition of Wodak, Kryzyanowski, Fairclough (1992) and many others who engage in studies of language use. Instead, it offers a way to critically interrogate discourses as knowledges, highlighted in Question 2 of the approach. It targets what it is possible to speak of by examining the knowledges (discourses) that limit and form “the sayable” (Foucault 1991: 59, 63). Rather than looking for “intended biases” or “rational argumentation”, the goal is to show how knowledges form their truths. This position is illustrated above in the description of how psychiatry as an accredited knowledge forms objects for thought – e.g., “behaviours” and “skills”. Approaching these concepts as “proposals about how we ought to proceed from here” best describes the WPR engagement with concepts.  

Relying on “context” 

Alongside the theoretically distinct approaches to discourse just described, it is important to reflect on how the tradition of discourse analysis as language analysis treats “contexts”. This topic was raised briefly in the previous entry (27 Feb. 2023) where we discussed how Koselleck (1998) appeals to “social history” as “occurring history” and “what ‘actually’ – and not just linguistically – happened”. There I describe his position as realist and, with Van Gelderen (1998), call for clarification of the ontological distinction between language and society. 

Whereas Koselleck lays out his position clearly, many theorists inadvertently adopt a realist perspective in relation to the conditions they identify as “background” in their analyses. I would imagine that many readers have a section on “background” in their work and I have also slipped into referring to “context” on occasion as if its meaning is straightforward. Happily, on one occasion, in the 2009 book I specify: “the task of ‘filling in’ context is never simply a descriptive exercise” since “reflections on context are themselves interpretive”. “Indeed”, I proceed to argue, “in many cases it is necessary to pay attention to how a context, for example, ‘globalisation’, is represented” (Bacchi 2009: 21). 

Genealogy in WPR (Question 3) acts as a corrective for or defence against the simple recording of “context/s”. Foucault explicitly identified the “role of genealogy” in tracing the “history of the concept of liberty or of the ascetic life”. In his view, the task is to make such conceptual histories appear “as events on the stage of historical process” (Foucault 1977: 1). “Events” in Foucault are not simply bits of data; they need to be analyzed in terms of the practices that give rise to them (Bacchi and Goodwin 2016: 46). These genealogies of practices demonstrate how selected concepts or categories of analysis (e.g., “alcohol problems”, “drug problems”) are made under specific circumstances, and hence, how they can be unmade (Bacchi 2015). 

In a special issue of Discourse & Society on “Theories and concepts in Critical Discourse Studies” Krzyzanowski and Forchtner (2016) explore some possibilities for theoretical developments within CDA and DHA. One contributor, Macgilchrist (2016), suggests that CDS (Critical Discourse Studies) must move beyond its theoretical and conceptual foundations and incorporate “post-foundational” thinking in discourse studies. She expressly endorses “productive integration” with the post-structuralist discourse theory associated with Ernesto Laclau (1990). 

At the same time Petersen (2015) cautions researchers who ascribe to poststructural precepts to be wary of the frequent tendency in their work to refer to social realities (“contexts”, “backgrounds”) as if they simply exist. In her view, “mainstream post-structuralist policy analysis is presently overwhelmingly realist”, while the serious and ongoing onto-epistemological challenges of post-foundationalism are “glossed over in order to ‘get on with it’” (Petersen 2015: 148).

It seems, then, that we all have some work to do. 

In the last of three entries on applying WPR to concepts, which follows in a month’s time, I pursue the concept of “crisis”, alongside “risk” and “uncertainty”. I hope to clarify how the distinct theoretical perspectives introduced in this and the preceding entry (27 Feb. 2023) engage these concepts. The task, using WPR thinking, is to discern how different usages of “crisis” involve contrasting proposals about how we ought to proceed from here (Tanesini 1994: 207).  


Bacchi, Carol 2000. Policy as Discourse: What does it mean? Where does it get us? Discourse: Studies in the cultural practices of education, 27(1), pp. 45–57.

Bacchi, C. 2005. Discourse, Discourse Everywhere: Subject “Agency” in Feminist Discourse Methodology’. NORA: Nordic Journal of Women’s Studies, 13(3): 198-209. Reproduced in C. Hughes (Ed.) (2012). Researching Gender. Sage Fundamentals of Applied Research Series. 

Bacchi, C. 2009. Analysing Policy: What’s the Problem Represented to be? Frenchs Forest: Pearson Education. 

Bacchi, C. 2015. Problematizations in alcohol policy: WHO’s “alcohol problems. Contemporary Drug Problems, 42 (2), 130–147.

Bacchi, C. and Bonham, J. 2014. Reclaiming discursive practices as an analytic focus: Political implications. Foucault Studies, 17: 173-192.

Bacchi, C. and Goodwin, S. 2016. Poststructural Policy Analysis: A Guide to Practice. NY: Palgrave Macmillan. 

Burr, V. 1995 An Introduction to Social Constructionism. London and New York: Routledge. 

Dean, M. 1999. Governmentality: Power and rule in modern society. London: Sage. 

Fairclough, N. 1992. Discourse and social change. Cambridge, MA: Polity Press.

Foucault, M. 1972. The Archaeology of Knowledge, translated by A. Sheridan Smith. New York: Pantheon Books. 

Foucault, M. 1977. Nietzsche, genealogy, history. In D.F. Bouchard, (Ed.), Language, counter-memory, practice: Selected essays and interviews. Ithaca: Cornell University Press.

Foucault, M. 1991. Politics and the study of discourse. In G. Burchell, C. Gordon & P. Miller (Eds), The Foucault Effect: Studies in Governmentality. Chicago: University of Chicago Press.

Koselleck, R. 1998. Begriffsgeschichte and social history. In I. Hampsher-Monk, K. Tilmans and F. van Vree (Eds) History of Concepts: Comparative Perspectives. Amsterdam: Amsterdam University Press. pp. 23-36.

Krzyzanowski, M. 2010. Discourses and Concepts: Interfaces and Synergies between Begriffsgeschichte and the Discourse-Historical Approach in CDA.  In R. de Cillia, H. Gruber, Krzyzanowski, M. and Menz, F. (Eds) Diskurs-Politik-Identitaet / Discourse-Politics-Identity. Tuebingen: Stauffenburg Verlag. 

Krzyzanowski, M. and Forchtner, B. 2016. Theories and concepts in critical discourse studies: Facing challenges, moving beyond foundations. Discourse & Society 2016, Vol. 27(3) 253–261.

Laclau, E. 1990. New Reflections on the Revolution of Our Time. London: Verso. 

Larner, W. 2008. Comments on Kevin Stenson’s “Governing the local: Sovereignty, social governance and community safety”. Social Work and Society – International Online Journal, 6 (1), 21–25, article/download/87/144. Accessed 14 January, 2016. 

Macgilchrist, F. 2016. Fissures in the discourse-scape: Critique, rationality and validity in post-foundational approaches to CDSDiscourse & Society, 27(3): 262–277. 

Petersen, E. B. 2015. What crisis of representation? Challenging the realism of post-structuralist policy research in education, Critical Studies in Education, 56:1, 147-160, DOI: 10.1080/17508487.2015.983940 

Potter, J. & Wetherell, M. 1990. Discourse: Noun, Verb or Social Practice? Philosophical Psychology, 3(2/3), pp. 205–218.

Reisigl, M. and Wodak, R. 2017. The Discourse-Historical Approach (DHA). In R. Wodak and M. Meyer (Eds) Methods for Critical Discourse Analysis. London: Sage. pp. 87-121. 

Tanesini, A. 1994. Whose language? In K. Lennon, & M. Whitford (Eds.), Knowing the difference: Feminist perspectives in epistemology. New York: Routledge. 

Van Gelderen, M. 1998. Between Cambridge and Heidelberg. Concepts, Languages and Images in Intellectual History.

In I. Hampsher-Monk, K. Tilmans and F. van Vree (Eds) History of Concepts: Comparative Perspectives. Amsterdam: Amsterdam University Press. pp. 227-238.Wodak, R. 2015. Critical Discourse Analysis, Discourse-Historical Approach. The International Encyclopedia of Language and Social Interaction, First Edition. Karen Tracy (General Editor), Cornelia Ilie and Todd Sandel (Associate Editors). © 2015 John Wiley & Sons, Inc. Published 2015 by John Wiley & Sons, Inc.
DOI: 10.1002/9781118611463/wbielsi116

Applying WPR to concepts: what’s to be gained?

In a paper that Anne Wilson and I delivered to the International Symposium held in Karlstad in August 2022, we introduce a way to apply WPR to concepts. The specific concept targeted is “underlying health conditions”. I have suggested in several places that it is possible to apply WPR to concepts (Bacchi and Goodwin 2016: 17). In the Symposium paper Anne and I show the usefulness of this form of application (Bacchi and Wilson 2022). 

Applying WPR to concepts involves approaching concepts as governmental problematizations. Such a theoretical stratagem shifts the focus from what concepts mean to how they are constituted and how they operate in governing practices. Recalling that the point and purpose of WPR is to interrogate how governing takes place – how we are governed (Bacchi 2009) – it is necessary, I argue below, to examine the role of concepts as governmental mechanisms. This position follows Dean’s (1999: 44-45) characterization of political language as a governmental technology. 

The Symposium paper emphasizes how the concept of “underlying health conditions” functions to shape political outcomes in specific circumstances. As examples, it examines the ways in which “underlying health conditions” as a concept, or category, impacts on consideration of COVID-19 death rates in several countries, and on health care options in the United States. In this way the example of “underlying health conditions” illustrates the WPR key premise that we are governed through problematizations ( KEYNOTE ADDRESS – CAROL BACCHI – 18 August 2022). It also highlights an important theoretical point – and one that causes endless confusion around WPR – that approaching concepts as governmental problematizations, in the manner just described, does not involve reflection on people’s assumptions or on people’s problematizations. We are working at a level of analysis distinct from considerations of intentionality. 

This Research Hub entry elaborates what WPR thinking adds to theoretical conversations about “concepts”, including Gallie’s (1955-56) “essentially contested concepts” and Koselleck’s (1982) “conceptual history”. It makes the case that applying WPR to concepts raises a wide range of issues about how governing takes place that do not appear in these other approaches to concepts. Through this discussion I take the opportunity to clarify the place of Foucault’s historical nominalism in WPR. In the next entry I consider how this discussion of “concepts” provides useful insights into how WPR differs from forms of Critical Discourse Analysis (CDA), specifically Discourse-Historical Analysis (DHA) (Wodak 2015; Reisigl and Wodak 2009; Krzyzanowski 2010). To forecast this argument, WPR is best described as offering an analysis of discourses (or knowledges) rather than as a form of discourse (language) analysis (Bacchi 2005). In the third entry in this series I use the concept of “crisis” to illustrate these arguments.

Thinking about concepts

What is a concept? How are we to identify concepts? There are numerous contributions to this topic. Despite the varied theoretical stances, there is some agreement that concepts are ambiguous in meaning and hence that attempts at definition are fraught. There are contrasting views about the meaning of ambiguity, however. The position that concepts are contested and/or contestable draws attention to the diverse interpretations associated with concepts such as “equality” or “agency” (Gallie 1955-56). As Hall (1998: 80) points out, this thesis does not take us “very far in any analytically useful direction”:

“For to claim that a particular concept is essentially contested, is to take an a-historical view of the character and function of political concepts. Not all concepts have been, or could be, contested at all times.”

Meanwhile, those who work in the field of “conceptual history” stress that concepts are ambiguous because they change meaning over time, again making definition a fraught exercise (Koselleck 1982). 

The Foucault-influenced poststructural stance adopted in WPR is a nominalist position where the terms used (e.g., “state”, “urban”) are simply the names assigned to “things” (Research Hub, 1 Oct 2020, 31 Oct 2020). Definition therefore is not simply a fraught exercise; it is an inappropriate one. The point of a poststructural analysis is not to defend one meaning over another but to consider how different meanings rely on contrasting presuppositions and lead to diverse effects. In tune with an ontology of becoming ( KEYNOTE ADDRESS – CAROL BACCHI – 18 August 2022) the focus is on the practices that produce concepts as “objects for thought”. You will hopefully recognize WPR thinking at work here. 

By turning to practices, Foucault looks to carve out a space between realism and idealism (Bacchi 2012: 3). In particular, he wishes to dislodge the impression of ideas or attitudes as some sort of objects whose history could be traced, a view sometimes expressed by those involved in “conceptual history” (Richter 1987).  Foucault sidesteps “ideas” and offers problematizations as a way to get inside “thinking”: 

“For a long time, I have been trying to see if it would be possible to describe the history of thought as distinct both from the history of ideas (by which I mean the analysis of systems of representation) and from the history of mentalities (by which I mean the analysis of attitudes and types of action [schemas de comportement]). It seemed to me that there was one element that was capable of describing the history of thought— this was what one could call the problems, or more exactly, problematizations”. (Foucault 1984)

Clearly, in WPR, a good deal of attention is paid to concepts and categories, which are themselves concepts, as outlined in the 2009 book (Bacchi 2009: 8-9). There it is mentioned that governing practices rely upon assumed categories, including people categories such as “youth”, “the homeless” and “citizens”, and conceptual categories such as “equality”, “discrimination” and many others. It follows that it is critically important to trace how these concepts and categories emerge and how they operate to shape political outcomes and people’s lives. In a previous entry I elaborate how this stance contrasts with “realist governmentality” (Stenson 2008) and with Critical Realism, where categories of analysis are taken to be markers of “real conditions” (Research Hub, 1 Feb 2019, 28 Feb 2021).

 In the broad area of “conceptual history” it is possible to draw distinctions between a Cambridge school and a German tradition described as Begriffsgeschichte (Conceptual History or, more precisely, The History of Concepts), associated with Koselleck (1982). Van Gelderen (1998: 230) outlines the differences in the two approaches. The Cambridge school embraces Pocock (1971) and Skinner (1989). It focuses on the history of political languages and originated in the Anglo-Saxon traditions of Collingwood (1946/1994), the philosophy of language, and J. L. Austin’s (1962) theory of speech acts. This “new history of political languages”, argues Van Gelderen (1998: 231), emphasizes “human agency as the prime mover of history, instead of the structures and processes floating through Begriffsgeschichte”. In the next Research Hub entry I show how this contrast can be overdrawn and how Begriffsgeschichte at times stresses the pronouncements of political actors, showing some affinity with the Cambridge school. 

The Begriffsgeschichte approach, overseen by Koselleck with his two co-editors Conze and Brunner, charts the “career of political and social concepts in German-speaking Europe between 1750 and 1850” (Richter 1986: 605). In the eight volumes that eventually were published (Koselleck 2011) and in other articles, Koselleck (1982) outlines the purpose and goal of this mammoth project, named the Geschichtliche Grundbegriffe [Historical Basics]. He argues that over the time period 1750 to 1850, which he calls the “Saddle period” or “Threshold period”, concepts became more abstracted, tending towards the form of what he calls “collective singulars”. Think here of concepts such as “liberty” or “equality”. As collective singulars, says Koselleck, these terms are available to be picked up and deployed to support ideologies and forms of propaganda (Richter 1986: 617). 

To make this argument, Koselleck (1998: 26) emphasizes the importance of social history, which he describes as “occurring history” and “what ‘actually’ – and not just linguistically – happened”. In this realist view, the “structures” and “processes” of social history exist separate from the linguistic structures of Begriffsgeschichte: “A history does not occur without speech, but it is never identical to speech and cannot be reduced to it” (Koselleck 1998: 26).  According to Van Gelderen (1998: 232), this ontological distinction between language and society “is in need of serious qualification”. I pursue this point in the next entry where I raise questions about the ontological status of “context”, which operates as “background” in much contemporary sociopolitical theory.

Following Foucault’s desire to carve out a space between realism and idealism, I borrow from the work of Tanesini to offer a poststructural approach to concepts – an approach that stands at a distance from both Koselleck’s “structures” and the emphasis on human agency at Cambridge. Tanesini (1994: 207) argues that concepts are not descriptive of anything; rather, they “are proposals about how we ought to proceed from here”. Their purpose is to “influence the evolution of ongoing practices”. Concepts do things; they are productive and can be described as performative ( KEYNOTE ADDRESS – CAROL BACCHI – 18 August 2022). In a sense then concepts are analogous to Austin’s (1962) illocutionary performatives that do not describe “reality” but that (help to) make worlds (Jackson 2004: 2). I emphasize “analogous” to avoid any confusion with the “speech act” position assumed by the Cambridge school (above).

Tanesini’s description of concepts as “proposals” provides a key link with WPR. In my work on WPR I argue that it is helpful to begin one’s analysis from proposals. This analytic strategy finds its rationale in the premise that what one proposes to do about something indicates what is targeted as needing to change and hence what is produced as “the problem”. Analytically, therefore, it is possible to start from proposals in texts and to work backwards to identify problem representations. In the previous two entries (30 Dec. 2022; 30 Jan. 2023) I provide pointers to how to identify “proposals” in governmental texts and other sources, including as examples buildings and political theories. Conceiving of “concepts” as “proposals about how we ought to proceed from here” provides the rationale for applying WPR to concepts.  

I offer several recent studies that illustrate the usefulness of treating concepts as problematizations.

  1. Sporer et al. (2018) explore the concept of “aspiration” as a technology of government in the UK and other OECD countries since the early 2000s – “Raising young people’s aspirations has been portrayed as a solution to persisting social and educational inequalities”. The authors ask how “aspiration” is constructed as a policy problem, paying particular attention to the genesis (genealogy) of the concept and how “youth” as a category is constituted both in terms of “deficit” and “potential”.
  2.  Lappalainen et al. (2019) analyze the “fluctuating” “conceptualisation of equality” in a cross-cultural analysis of Finnish and Swedish upper secondary curricula from 1970 to 2010s. They apply the WPR questions to selected policy texts, paying particular attention to the “problem of equality” and “the presuppositions that underlie the problem as presented”. 
  3. Miranda-Molina (2023) examines how the “problem” of “leveling” is represented in Latin American higher education policy documents. He offers useful insights into how the concept of “leveling” (nivelación) creates the “problem” of the lower retention rates of underrepresented social groups as “a consequence of their insufficient preparation”.

In each of these three cases the authors effectively probe the specific uses of concepts to identify their governing effects. In each case they approach concepts, not as taken-for-granted “truths”, but as “proposals about how we ought to proceed from here”. 

This discussion takes us to the critical issue raised earlier where I noted that approaching concepts as governmental problematizations does not involve reflecting on people’s assumptions or on people’s problematizations. Rather, the problematizations found in the proposals associated with any particular concept ought to be viewed as governmental mechanisms not as the products of political manipulation or rhetorical deployment. I hope that these three examples illustrate this point clearly. This theme is pursued in the next entry. 


As I suggested at the outset, I believe that bringing WPR to concepts allows us to raise questions that do not get raised in other approaches to concepts. Specifically, through asking how a particular concept problematizes an issue, we can bypass the question of intentionality and get inside thinking. Considering the effects of specific problematizations, meanwhile, allows us to see how concepts operate as governing mechanisms.

Returning to where we started, while preparing our paper, Anne Wilson and I reflected on the way in which “underlying health conditions” could be seen to support a particular political agenda. On the one side the designation of some COVID-19 deaths as associated with “underlying health conditions” seemed to suggest almost deliberate malfeasance on the part of our politicians and public health spokespeople who sought a rationale to reduce restrictions on public movements and trade. On the other side the determination of the WHO to “name” all COVID-19 related deaths as due to COVID-19 appeared to suit their political goal of insisting on the urgency of increased public health restrictions. 

However, the point of our Symposium paper is not to identify the place of problem representations in political manoeuvring of this sort. Rather, we wanted to understand how this category of “underlying health conditions” became a taken-for-granted way of thinking among health officials and the general public – and how it sets the scene for how we were/are governed. This story involves the long historical process of “making” distinctions in health/illness categorization – e.g., “acute” versus “chronic” illness – and the development of the concept of “co-morbidity”. Such categories emerged from the introduction of public health records and monitoring of population trends. 

The issues we face around the uses of “underlying health conditions”, therefore, emerge from multiple heterogeneous practices rather than from explicit manipulation for political ends. They are issues to do with how we conceive of “health” and “illness”, and how these shape “the evolution of ongoing practices” (Tanesini 1994: 2007). It is this broad governing agenda that WPR targets through the analysis of “concepts” as problematizations. The importance of distinguishing this position from the popular view in much political and policy analysis that our analytic focus should be people’s problematizations provides the starting point for the next entry.


Austin, J. L. 1962. How to do things with words. Massachusetts: Harvard University Press.

Bacchi, C. 2005. Discourse, Discourse Everywhere: Subject “Agency” in Feminist Discourse Methodology’. NORA: Nordic Journal of Women’s Studies, 13(3): 198-209. Reproduced in C. Hughes (Ed.) (2012). Researching Gender. Sage Fundamentals of Applied Research Series. 

Bacchi, C. 2009. Analysing Policy: What’s the Problem Represented to be? Frenchs Forest: Pearson Education. 

Bacchi, C. 2010. Why study problematizations? Making politics visible. Open Journal of Political Science, 2(1): 1-8.

Bacchi, C. and Goodwin, S. 2016. Poststructural Policy Analysis: A Guide to Practice. NY: Palgrave Macmillan. 

Bacchi, C. and Wilson, A. 2022. Governing through “underlying (preexisting) health conditions”: “chronic illness”, “race-ism” and COVID-19. Paper presented at the International Symposium on Critical Policy Studies, Karlstad University, 17-19 August 2022. Paper available on request from C. Bacchi –

Collingwood, R. G. 1994. The Idea of History. Oxford: OUP. First published in 1946. 

Dean, M. 1999. Governmentality: Power and rule in modern society. London: Sage. 

Foucault, M. 1984. Polemics, Politics and Problematizations; based on an interview conducted by Paul Rabinow. Translated by Lydia Davis. Appears in Essential Works of Foucault, Vol 1: Ethics. New Press.

Gallie, W. B. (1955-56) Essentially Contested Concepts. Proceedings of the Aristotelian Society, 56: 167-98.

Hall, T. 1998. Conceptual History and the History of Political Thought. In I. Hampsher-Monk, K. Tilmans and F. van Vree (Eds) History of Concepts: Comparative Perspectives. Amsterdam: Amsterdam University Press. pp. 75-86.

Jackson, S. 2004. Professing Performance: Theatre in the Academy from Philology to Performativity. Cambridge: Cambridge University Press.

Koselleck, R. 1982. Begriffsgeschichte and social history, Economy and Society, 11:4, 409-427, DOI: 10.1080/03085148200000015 

Koselleck, R. 1998. Begriffsgeschichte and social history. In I. Hampsher-Monk, K. Tilmans and F. van Vree (Eds) History of Concepts: Comparative Perspectives. Amsterdam: Amsterdam University Press. pp. 23-36.

Koselleck, R. 2011. Prefaces to volumes 3, 6, 7, and 8 of the Geschichtliche GrundbegriffeContributions to the History of Concepts, 6(1). pp. 25+. Gale Academic OneFile, Accessed 5 Jan. 2023.

Krzyzanowski, M. 2010. Discourses and Concepts: Interfaces and Synergies between Begriffsgeschichte and the Discourse-Historical Approach in CDA.  In R. de Cillia, H. Gruber, Krzyzanowski, M. and F. Menz (Eds) Diskurs-Politik-Identitaet / Discourse-Politics-Identity. Tuebingen: Stauffenburg Verlag. 

Lappalainen, S., Nylund, M. and Rosvall, P. 2019. Imagining societies through discourses on educational equality: A cross-cultural analysis of Finnish and Swedish upper secondary curricula from 1970 to the 2010s. European Educational Research Journal 2019, 18(3) 335–354 

Miranda-Molina, R. 2023. Preuniversitarios, ciclos iniciales y apoyos suplementarios: políticas latinoamericanas de “nivelación” y sus problemas representados. Revista Dilemas Contemporáneos: Educación, Política y Valores, 2(4).

Pocock, J. G. A 1971. Politics, Language and Time: Essays on Political Thought and History. Chicago: University of Chicago Press. 

Reisigl, M. and Wodak, R. 2009. The Discourse-Historical Approach (DHA). In R. Wodak and M. Meyer (Eds) Methods for Critical Discourse Analysis.  London: Sage, pp. 87-121.

Richter, M. 1986. Conceptual History (Begriffsgeschichte) and Political Theory. Political Theory, 14(4). 604-637. 

Richter, M. 1987. Begriffsgeschichte and the History of Ideas. Journal of the History of Ideas, 48(2): 247-263.  

Skinner, Q. 1989. Language and political change. In T. Ball, J. Farr and R. L. Hanson (Eds) Political Innovation and Conceptual Change. Cambridge, UK: Cambridge University Press. pp. 6-23.  

Spohrer, K., Stahl, G. & Bowers-Brown, T. 2018. Constituting neoliberal subjects? ‘Aspiration’ as technology of government in UK policy discourse, Journal of Education Policy, 33:3, 327-342, DOI: 10.1080/02680939.2017.1336573 

Stenson, K. 2008. Governing the Local: Sovereignty, Social Governance and Community Safety. Social Work & Society, 6(1).

Tanesini, A. 1994. Whose language? In K. Lennon, & M. Whitford (Eds.), Knowing the difference: Feminist perspectives in epistemology. New York: Routledge. 

Van Gelderen, M. 1998. Between Cambridge and Heidelberg. Concepts, Languages and Images in Intellectual History.

In I. Hampsher-Monk, K. Tilmans and F. van Vree (Eds) History of Concepts: Comparative Perspectives. Amsterdam: Amsterdam University Press. pp. 227-238.Wodak, R. 2015. Critical Discourse Analysis, Discourse-Historical Approach. The International Encyclopedia of Language and Social Interaction, First Edition. Karen Tracy (General Editor), Cornelia Ilie and Todd Sandel (Associate Editors). © 2015 John Wiley & Sons, Inc. Published 2015 by John Wiley & Sons, Inc.
DOI: 10.1002/9781118611463/wbielsi116

WPR: Starting from scratch

Given the new year is upon us, it seems timely to reflect on the possibilities and challenges involved in starting a WPR analysis from scratch. 

I ask you to imagine that you have just encountered the WPR approach for the first time, and you are wondering how it can be of use to you in your selected field or with the specific topic that has attracted your interest. 

I feel better able to engage with this question at the moment as I have recently been involved in exactly this exercise – asking myself what it means to apply WPR to a new field of interest.  I will share more of these details in several months’ time once the project is completed.

To begin, it is necessary to reflect on whether or not WPR is suited to the project you have in mind. What kind of analysis is WPR intended to undertake and does this kind of project fit your goals? 

WPR provides a means to build up an understanding of, and to interrogate, how governing takes place. Importantly, it adopts a view of governing that embraces more than conventional political institutions. The goal is to develop a fine-grained picture of the complex and intermingled factors and forces shaping lives. A particular emphasis is placed on the “knowledges” involved in governing, and hence on the role of experts and professionals. So, for example, it becomes important to think about the ways in which premises from psychology and related fields shape governing mechanisms. The place of behavioural economics in “nudge theory” (Research Hub entry, 26 Nov. 2017) offers an example. Note that WPR does not operate at the level of people’s assumptions – if this is what interests you, you need to find a different analytic framework. In WPR the goal is to identify and interrogate the epistemological and ontological assumptions required to give a specific policy (read broadly) meaning. 

As we saw in the last entry (30 December 2022) it is possible to start one’s analysis from a piece of legislation or government report, and to show how it relies upon deep-seated presuppositions (“knowledges” such as behavioural economics) to make sense. I talk there about using pieces of legislation as “levers” to open up governing practices to critical questioning (Bacchi and Goodwin 2016: 18, 20). Another way to say this is to say that WPR looks to open up and examine the “space being governed”, or the “problem-space” (Walters 2004: 247). To this end we explore governmental problematizations.

Here it is useful to remember that, for WPR, “government”, in the broad sense just described, is best approached as a “problematizing activity” (Rose and Miller 1992: 181). Osborne (1997, p. 174) concurs that “policy cannot get to work without first problematizing its territory.” That is, in order for something to be governed, or imagined as governable, it needs to be problematized (Packer, 2003, p. 136).  Problematizations therefore provide a useful starting place to reflect on how governing takes place. 

Your first objective therefore is to see how your selected topic of interest is being problematized. The fact that something is the target of legislation – think for example of the National Youth Suicide Prevention Strategy 1995/97 – means that the situation or condition is being problematized in a particular way. In Analysing Policy (2009: 4), I show how the approach to youth suicide as a “problem” in the Prevention Strategy encompasses psychologists, parents and researchers. You can see here how the notion of “governing” is broadened beyond the legislative instrument to include a wide-ranging array of groups, and also how it becomes useful to ask what kind of “problem” “youth suicide” is represented to be. 

Once you determine that your area/topic of interest is usefully approached through the lens of problematization, you face three tasks:

  1. Expand your understanding of the history, background and “context” of your selected area of interest.
  2. Select specific proposals to gain access to the problematizations at work.
  3. Apply the WPR questions to your identified problematizations.

I will run through each of these tasks, outlining what they entail.

One last and important introductory point needs to be made. To say that governing takes place through problematizations does not mean that these ways of conceptualizing an issue are automatically adopted and/or effective. Miller and Rose (1990: 10) describe “government” as a “congenitally failing operation”, requiring continuous and repeated efforts to shape citizen behaviours. At the same time, it is important to consider the effects problem representations have on people and practices, undertaken in Question 5 of WPR (see Bacchi and Goodwin 2016: 20). For example, we can reflect on the influence of “nudge theory” on specific policy areas and how “nudges” may shape people’s behaviours (they are certainly aimed at this outcome). This stance retains space for both resistance and contestation. 

Task 1: History, background and “context”

Starting a WPR analysis from scratch entails a good deal of reading and research. Usually, you choose a topic because it interests you and you probably already have some background. Still, it is useful to approach the task of building one’s understanding of the situation systematically. I suggest developing a “web of policies”, and allied texts, to show how your selected topic fits into a larger picture or pictures. This web will involve you necessarily in thinking of the long-term development of specific policy interventions. Here, I stress the importance of a “genealogical sensibility” (Research Hub entry 30 July 2022). In genealogy we are not looking for causes but for linkages, and we are not tracing path dependence but acknowledging contingency. This task can be wide-ranging and can take you in directions you had not anticipated. It is important to keep an open mind about possible connections among particular developments. In addition, the task of “filling in” context is not a descriptive exercise; instead, it is necessary to pay attention to how “contexts” are themselves represented (Bacchi 2009: 21).

The task of exploring “context” extends in a second direction – to reflect on connections among aspects of your selected topic and wide-ranging philosophical perspectives. Why philosophical perspectives, you may ask? Because invariably stances on political issues involve views on a range of related topics – e.g., the usefulness or not of education, the purpose of migration policies, the meaning of equity in relation to equality, and so on. These topics – which are offered as examples – are grounded in competing pedagogical philosophies, conceptions of human justice, and meanings of equality. 

In Analysing Policy (2009: 21, 56) I suggest that the concept of “nesting” may assist you in dealing with these complex connections. The point here is to recognize that any policy will necessarily intersect with specific views on related philosophical issues. To deal with Question 2 in WPR on epistemological and ontological assumptions requires that we reflect on these intersections and how they influence the selected topic or topic area. 

In Women, Policy and Politics (Bacchi 1999: Chapter 6), for example, I show how competing approaches to reform in the area of girls and education are grounded in competing problematizations of educational practices. If one adopts a critical stance on education – seeing it as more oppressive than emancipatory – it is unlikely that a researcher will endorse proposals to increase women’s representation in higher education institutions as a means to promote “women’s equality”. “Nesting” therefore alerts you to the need to ask the WPR questions at several stages of the analysis. In this case it would be necessary to ask, “What is the ‘problem’ of ‘education’ represented to be?” and also “What is the ‘problem’ of ‘equality’ represented to be?” 

In terms of “context”, it is also important to remember that WPR aims to assist us to understand patterns in forms of governmental problematization, which are described as “styles of problematization”. Basically, problematizations allow access to the “thinking” in modes of rule. They do this through a focus on the rationales (or rationalities) offered for specific modes of rule and through examining how specific governmental technologies operate – the means by which governing becomes practicable. By examining the “problem-space” in governing practices, we can identify the logics/rationalities at work and place them under scrutiny. It is important here not to enshrine a particular mode of rule, such as neo-liberalism, as some sort of ideal type or determining influence (see Larner 2000). 

Finally, as you prepare yourself to see what WPR can bring to your analysis, it is important to read widely in the literature on the topic. I emphasize in particular the need to seek out critical literature. It can be difficult for researchers to perceive the impact of accepted frames of reference on their analysis and critical literature can stimulate fresh perspectives.  

Task 2: Select specific proposals to gain access to the problematizations at work.

With our broadened background, the next task becomes finding proposals or proposed solutions that engage issues that you deem to be relevant. To this end we examine what Foucault calls “practical texts” or “prescriptive texts”, which provide guides to conduct (Bacchi 2009: 34). 

WPR starts from the premise that what one proposes to do about something indicates what is targeted as needing to change and hence what is rendered problematic, or “the problem” (Bacchi 2009: 2-3). This simple premise provides researchers with guidance on how to approach their selected text/s. But what are proposals? And how are we to identify them? 

Proposals can take several forms. Often a selected text will have recommendations within it, and it is fairly clear that recommendations for change are proposals for change (see also “aims”). But the process of problematization is much more nebulous than this example suggests. If a text, for example, praises initiatives aimed at developing (more) social cohesion, you can read this comment as a problematization, proposing the need to increase social cohesion (i.e. the “problem” is represented to be inadequate social cohesion). If you are looking for key terms that may provide assistance in identifying proposals, I find the word “should” a useful option. When a text suggests that something should be done (see also “must” and “shall”), it can often be read as a proposal to achieve a specific goal. 

It needs to be remembered that you are looking for proposals simply to provide a focus for the remaining WPR questions. They are a way “in” to your text of choice. I often use the example of training programs for women as a reform targeting the goal of increasing women’s representation in positions of influence. As I say: if training programs are the proposal, the problem is represented to be women’s lack of training (Bacchi 2009: x). So, I talk about starting from the proposal/s and “working backwards” to identify the problem representation, OR about “reading off” the implicit problem representation/s from the proposal/s.

I’m asked if this sort of thinking produces WPR as necessarily negative in its thinking. In contrast, I argue that women’s “lack of training” serves only as a starting point for your analysis. It helps to open up the topic area or “problem space” in useful ways. And it does this without imposing an interpretation on the issue under consideration. The problem representation emerges from the text itself. 

Task 3: Apply the WPR questions to your identified problematizations or problem representations.

The most recent version of the WPR questions is available in Bacchi and Goodwin 2016, p. 20. The questions are challenging to apply primarily because they rely upon a range of theoretical perspectives, including perspectives on power, social change, conceptions of the subject, and so on. Chapter 3 in Bacchi and Goodwin offers a basic introduction to these concepts.  

Question 2 brings to attention some underlying premises and presuppositions (conceptual logics) that help to make the identified problem representation/s intelligible. It’s helpful to think about the governing “knowledges” (e.g., behavioural economics) mentioned above and their pivotal contribution to governing practices. You can see here that the focus in this form of analysis is not on producing “truth” but on interrogating the mechanisms that produce something as “true” or as “in the true” (Foucault 1991: 58). 

Question 3 invites a genealogy of an identified problematization. How did this representation of the “problem” come to be? In WPR there is no search for ultimate causes; rather, the emphasis is on contingency and heterogeneity. 

Question 4 asks what is not problematized in this particular problem representation. The goal here is to broaden the conversation and to draw attention to aspects of the issue that have been ignored. Researchers use this question to search out critical perspectives that deserve reflection.

Question 5 targets effects, or implications. I talk about effects under three headings: discursive, subjectification and lived. Most recently I have suggested that objectification also needs to be explicitly included ( KEYNOTE ADDRESS – CAROL BACCHI – 17 August 2022).

Question 6 invites more attention to the specific practices that install particular knowledge regimes and problematizations. It also specifies the need to seek out signs of resistance.

Step 7 draws attention to the key importance of self-problematization. The point here is to recognize that every researcher is embedded in specific knowledge regimes and hence there is every likelihood that they could buy into assumptions and presuppositions that require interrogation. Whereas many research fields refer to the need for “reflexivity”, I suggest that we need to conceive of this self-questioning practice as a practice of the self – actually applying the WPR questions to one’s own proposals. 

Importantly, WPR is not a formula. Moreover, the questions are clearly interconnected. Still, many researchers find it useful to adopt the list of WPR questions to structure their argument. PhD and Masters’ students often find this approach helpful. However, it is also possible simply to allow the questions to operate in the background of an analysis, without addressing each question separately. I refer to this form of application as an integrated analysis. Examples of how to produce integrated analyses are available in Bacchi 2009, Chapters 5 through 10.

I am asked if some of the WPR questions can be omitted, and others targeted. I can see why this proposition may appear necessary given the complexity involved in producing a genealogy, for example. Since I have just made the point that the questions can operate in the background and may not need mentioning at all, it would appear that foregrounding certain questions is feasible. At the same time the WPR questions form part of a way of thinking, a way of thinking reliant on a range of epistemological and ontological assumptions (see Bacchi and Goodwin 2016, Chapter 3). Indeed, it would be possible and could be useful to apply WPR to WPR in order to open up these grounding premises to critical interrogation. I will consider such a project in the near future. 

In the meantime, enjoy!


Bacchi, C. 1999. Women, Policy and Politics: The Construction of Policy Problems. London: Sage.

Bacchi, C. 2009. Analysing Policy: What’s the Problem Represented to be? Frenchs Forest: Pearson Education.

Bacchi, C. and Goodwin, S. 2016. Poststructural Policy Analysis: A Guide to Practice. NY: Palgrave Macmillan. 

Foucault, M. 1991. Politics and the study of discourse. In G. Burchell, C. Gordon, P. Miller (Eds.), The Foucault effect: Studies in governmentality. Chicago: University of Chicago Press.

Larner, W. 2000. Neo-liberalism: Policy, Ideology, Governmentality. Studies in Political Economy, 63: 5-25.

Miller, P., & Rose, N. 1990. Governing economic life. Economy and Society, 19 (1), 1–31. 

Osborne, T. 2003. What is a problem? History of the Human Sciences, 16, 1–17.

Packer, J. 2003. Disciplining mobility: Governing and safety. In Z. Bratich, J. Packer, & C. McCarthy (Eds.), 

Foucault, cultural studies and governmentality (pp. 135–161). New York, NY: State University of New York Press.

Rose, N., & Miller, P. 1992. Political power beyond the state: Problematics of government. British Journal of 

Sociology, 43, 172–205.

Walters, W. 2004. Secure borders, safe haven, domopolitics, Citizenship Studies, 8:3, 237-260, DOI: 10.1080/1362102042000256989