By Rick Szostak
Unknown unknowns are challenges that we will face in future that we do not foresee today.
Here I argue that an important subgroup of unknown unknowns occurs when some phenomenon that we know a lot about has an unexpected effect on another phenomenon that we know a lot about, especially when there are few links between the two silos of knowledge. An example is unanticipated “interactions” between medications prescribed by medical practitioners from different specialities. Here I explore such disciplinary interactions more generally.
Disciplinary scholars focus on interactions among the phenomena that their discipline studies, but usually ignore interactions with phenomena studied in other disciplines. The academy as a whole thus devotes little attention to interactions among phenomena studied in different disciplines.
I have explored major historical transformations, which were generally surprises at the time and found they always involve interactions among the phenomena studied by multiple disciplines. In Szostak (2021) I provide flowcharts of dozens of such transformations, all involving cross-disciplinary interactions. Just two examples are:
- the development of agriculture which arguably reflected (at least) changes in population, environment, and technology and, in turn, had dramatic effects on environment, population, economy, politics, and social structure.
- the expansion in the size of state bureaucracies over the last century which reflects political, economic, and technological developments, and has impacts on phenomena studied across the human sciences.
I argue that it is a general historical rule that major transformations involve cross-disciplinary interactions, and that this insight provides a powerful guide to understanding the future.
Further, I argue that most disciplines posit some system of stability among the phenomena they study (Szostak 2017). Yet, these systems of stability are often disrupted by interactions with phenomena studied in other disciplines. For example, the market equilibria studied by economists can be disrupted by changes in weather or public tastes or political institutions.
Here we have a very important sort of unknown unknown: a system that generally exhibits stability – and thus we may not worry about it too much – is suddenly shocked into dramatic change by factors considered exogenous by disciplinary experts.
Although disciplinary scholars are aware that their systems can be shocked from outside, they devote far less attention to exogenous shocks than to theorizing systems of stability.
At other times, they may purposely ignore or downplay the possibility of cross-disciplinary interactions. For example, economists know that there was a Great Depression in 1929-1939, but happily develop macroeconomic theories that cannot explain it, and, as I showed in Szostak (2005), are hostile to explanations of the Great Depression that involve phenomena studied in other disciplines.
The crux of my argument is that disciplinary silos generate unknown unknowns.
It seems logical to posit, therefore, that cross-disciplinary scholarship is a way of identifying and mitigating unknown unknowns. This opens important new directions for cross-disciplinary scholarship and raises a key question:
How can we best study cross-disciplinary interactions so as to know about unknown unknowns?
In your own work:
- Have you seen examples in your own discipline of a reluctance to embrace useful ideas from outside?
- Do you have instances where cross-disciplinary research uncovered unknown unknowns?
Szostak, R. (2005). Evaluating the Historiography of the Great Depression: Explanation or Single-Theory Driven? Journal of Economic Methodology, 12: 1, 35-61.
Szostak, R. (2017) Stability, Instability, and Interdisciplinarity. Issues in Interdisciplinary Studies, 35: 65-87. (Online – open access): http://interdisciplinarystudies.org/docs/Vol35_2017/05_65-87.pdf (PDF 250KB).
Szostak, R. (2021). Making Sense of World History. Routledge: London, United Kingdom. (Online – open access; e-book published October 2020) (DOI): https://doi.org/10.4324/9781003013518.
Biography: Rick Szostak PhD is Professor and Chair of Economics at the University of Alberta, Canada. His research interests span the fields of economic history, world history, interdisciplinary studies, knowledge organization, and future studies. His book Making Sense of the Future will be published by Routledge in 2021. He coordinated the development of the About Interdisciplinary and Interdisciplinary General Education pages listed under resources on the Association for Interdisciplinary Studies website.
This blog post is part of a series on unknown unknowns as part of a collaboration between The Australian National University and Defence Science and Technology.
For all blog posts published in this series, see: https://i2insights.org/tag/partner-defence-science-and-technology/
23 thoughts on “Can examining cross-disciplinary interactions illuminate unknown unknowns?”
Hi Rick, thank you for this valuable article. The consideration of unknown unknowns is an important aspect of knowledge management (KM), so I’m very pleased to share your article in RealKM Magazine at https://realkm.com/2021/02/22/can-examining-cross-disciplinary-interactions-illuminate-unknown-unknowns/
I very much agree with your proposition, based on my own experience of where research went beyond disciplinary silos and in doing so uncovered unknown unknowns. I’ve document one such case study of this in the article at https://realkm.com/2016/01/19/dangers-information-silos/
I then further explore the disciplinary silo issues that were the basis for the issue discussed in that case study in the follow-on article at https://realkm.com/2019/10/30/case-study-how-polarized-debates-can-be-the-result-of-rational-deliberation-and-how-they-can-be-resolved/
The second article includes discussion of approaches to bridging disciplinary boundaries from a 2019 paper.
Thanks Bruce!; I really appreciate the re-posting.
And I like your argument about how polarization can reflect access to different pools of information, and thus can be reduced by knowledge transfer. This is an important insight that I have not seen framed in this way before.
Thanks again, Rick
Hi Rick, thanks for providing your great article for discussion. I particularly like how you have framed this discussion with your title as a question, which, in my view, sets us up with precisely the right outlook about the topic.
I also had a quick look at your referenced article; the development of agriculture as an example piqued my interest just because I’ve been reading a little about both this and the changing views about the Bronze Age Collapse, wherein recent views seem to favour a complex interaction of events rather than just “the sea peoples did it”. While you have indeed, as advertised, cast the issue in terms of gaps in knowledge across disciplines or fields, I think there is plenty of evidence that the same mechanics is well at work within fields as well across them. An obvious example would be the somewhat traumatic conflict between General Relativity and Quantum Dynamics (the conflict between General Relativity and Newtonian Dynamics is not talked about in such terms because we well understand the latter as an approximation within well understood bounds of the former). I also don’t think that it is necessarily the case that intra-disciplinary manifestations of incommensurability between theories are inherently subordinate in scale or difficulty or importance to inter-disciplinary manifestations, either. I suggest this is an important part of the story, because it highlights that a significant difference is that conflicts within a notional field or discipline are generally well recognised, while those between fields or disciplines may generally be not recognised. Might this even be the delineating criteria for what constitutes a ‘field’ or ‘discipline’? Might we identify two fields or disciplines if the relationships between the component topics or theories are not well understood?
I suppose my thought really boils down to consequences of the logical inevitability of the incommensurability of different but individually successful theories – which I would define in terms of reliable solutions to equally provisional problem formulations – with the multi-discipline opportunities you’ve so nicely described and with which we’re probably all deeply engaged then being perhaps better viewed as driving the need for better problem formulations rather than just alignment of solutions.
I agree that there can sometimes be substantial differences within disciplines. Some disciplines have fields that interact little: cultural anthropologists and physical anthropologists have little overlap in theories or methods, for example. So you are entirely correct that the problem I identify can occur within disciplines.
Yet there are big institutional reason that differences within disciplines tend to be smaller than differences across disciplines. Most hiring at research universities is done by disciplinary-based departments. Also, teaching assignments at research universities are also made at the department level. There are often introductory courses that attempt to survey the discipline. So researchers in disciplines get to police disciplinary boundaries, and have incentives to have a general understanding of the discipline as a whole. My economist colleagues all know more about some fields in economics than others but generally know all of these better than they understand other disciplines.
When there are theoretical disputes within a discipline, good scholars should try to understand those they disagree with. But there is no expectation in the academy that scholars be familiar with theories in other disciplines.
I tend not to use the word “incommenurability” I think we always have some ability to understand each other, and compare and contrast different theories, and that there are strategies we can employ to transcend differences in terminology and perspective. So I think the greatest danger is when one group of scholars just doesn’t interact with another group. Then it is particularly likely that the phenomena that one group studies has unanticipated effects on the phenomena that the other group studies.
Thanks for the important clarification, Rick
Hi Rick, thanks again. Nice point about incommensurability: I definitely intend this differently to you, so this is an important thing I’ll keep in mind: I mean it in a formal logical sense whereby the framing basis of theories manifests a fundamental incompatibility, such as occurs between logic systems that are individually valid but that have incompatible semantics. Playing with this in terms of theories to capture complex relationships between other theories and hence about contextuality is daily business in one of my favourite collaborative projects. The same thing is at the heart of the incompatibility between Quantum Dynamics and General Relativity: they are predicated on logically incompatible constructions. Yet plenty of people in these areas most definitely understand both, so it isn’t an incompatibility between researchers, just between theories. I suppose what I’m thus driving towards is that we ought to be aiming at dissolving (or resolving?) the apparent framework and language barriers between researchers across fields to the point where what we are left with is the beautiful logical incommensurabilities between theories as our primary objects of collaborative study. From experience, I would propose that achieving this is what all the hard work of building successful cross-disciplinary collaborations is really all about.
Thanks again Darryn; I agree entirely that we need to be dissolving barriers between different groups of scholars. I am probably more optimistic than you that we can then seek to integrate differing points of view. But for the purposes of identifying unknown unknowns the key step is facilitating communication across different groups of scholars.
Thanks for your contributions, Rick
Hi Rick, you’re welcome, and this has been a most interesting and useful discussion for me – it’s helped me to position some nagging observations from experience in building multi-disciplinary teams much better. I’m deeply grateful to you for this. I’d probably only just say here that, to me, identifying logical incommensurabilities between theories across fields is a pinnacle of optimism, because it leads not so much to the integration of existing ideas as to the generation of much better problem choices: this is expressly a primary goal that researchers in a very active research community that is otherwise unfortunate enough to have me as its facilitator explicitly pursue. As this community has matured, integrating existing ideas has become almost disappointing by comparison to using incompatible ideas from across fields to then create new candidate problem conceptions to study; thus researchers noticeably get excited and celebrate it when we tease out why and how different ideas representing distinct perspectives just won’t fit together. I most certainly agree with you that facilitating the communication across groups to get to this point of maturity is a huge effort, one that organisations appear universally at present to vastly underestimate. Thanks again!
Hi Darryn; It has been interesting and useful for me too.
If some of the straightforward strategies for integration don’t work, then researchers need to examine the assumptions of competing theories. So I think that the process you describe might be thought of as a way forward to achieve some sort of consensus in the most difficult cases of theory conflict. Integration is a process that sometimes takes a very long time.
But perhaps I am stretching the definition of integration to the breaking point.
Thanks for your insights. I suspect that researchers quite enjoy being facilitated by you, Rick
Thank you for sharing your thoughts about unknown unknowns and silos, Rick. This week I have been reading in two related areas that may interest you:
a. cynefin framework which uses the same term for complexity (e.g., https://www.thinknpc.org/resource-hub/how-to-translate-decision-into-action/)
b. dialogic organization development which presents a framework for creating meaning in complex situations through facilitated discourse (e.g., https://b-m-institute.com/wp-content/uploads/2019/10/ODP-V45No1-All_Pages.pdf)
At this point in my growing understanding there seem to be at least two sources of the silo effect:
1. lack of a shared language, as with disciplines
2. the categorization of the problem by one group as simple or complicated rather than complex and by the other group at a higher level.
I will look forward to your next book.
Thanks Mike! I very much like the way that we are collectively seeing connections to literatures in many fields.
Your closing remarks remind me of O’Rourke, M., Crowley, S., Eigenbrode, S. D., Wulfhorst, J. D., eds. Enhancing Communication and Collaboration in Interdisciplinary Research. Thousand Oaks: Sage, 2013, It identifies two main barriers to collaboration: differences in terminology and differences in perspective. Different chapters suggest strategies for addressing each.
There are also institutional barriers. I wonder if dialogic OD could encourage interdisciplinarity within universities?
Thanks again, Rick
Thank you for the insights, Mike. I would suggest that metrics can also be a force for silos — being measured in different ways can create a blind spot to other aspects of operations or planning
Thanks for the interesting blog, Rick, and congratulations with the ambitious & very well readable world history book! This ‘unknown unknowns’ perspective on disciplinary respectively interdisciplinary knowledge is certainly insightful. Yet I wonder whether this is just an epistemological issue (as presented here) or whether it also has an ontological dimension. Objects as diverse as planets and human animals have metaphysical and functional properties that appear to be independent of the properties of other objects yet at the same time these properties are often related to other -similar and dissimilar- objects. For example, the chemical and physical properties of such objects have of course a causal/etiological history which connects them to others all the way to the big bang. Yet such connections are even closer to home as when their weight and temperature and chemical activity is dependent upon gravity and radiation, caused by the presence of other planets and stars. Such causal connections are more impactful for animals, of course. So is independence / dependence of objects from others (and hence also stability versus interactive dynamics) not always a matter of degree, both epistemologically and ontologically speaking? With interdisciplinarity being particularly relevant for articulting their interdependence & dynamics?
Thanks Machiel: I think that our epistemological reality is to some extent a reflection of an ontological reality. We have economics departments because there is an economic system(s) in which a set of economic phenomena interact. And we have chemistry departments to study how chemicals interact under certain circumstances. What we miss in our disciplinary set-up is the ontological reality that all of these smaller systems interact (in what John calls a system-of-systems). It perhaps made sense for humans to first try to understand these smaller systems but we now face a range of environmental and social problems that require us to cope with bigger systems than any one discipline can engage.
So I would say that the essence of our ontological reality is “everything affects (almost) everything else.” But our epistemological reality is “we mostly study how some things affect a narrow set of other things.”
Thanks again, Rick
Although I agree in a general sense with your statement that “the essence of our ontological reality is “everything affects (almost) everything else””, I do think that there are differences in degree in these interactions. I would contend that generally speaking objects tend to have maximal interactions with objects that function at the same level of organization whereas cross-organizational-level interactions are in many cases explanatory less relevant – also because these are in turn mediated by objects at inermediate levels. E.g. the impact of quantum phenomena on our cognition or of gravitational wave ripples on our economies is explanatorily irrelevant or insignificant, since they can’t help to explain variations in our cognitive or economical phenomena. Nonetheless, although I think there is some reason for our disciplinary ‘epistemological reality’, what is lacking is a general awareness how & when inter-level interactions might turn out to be explanatory more relevant in certain cases than usually and how to approach these in an interdisciplinary way.
Thanks Machiel; You take our conversation to the next level! (pun intended)
How indeed do we identify the cross-disciplinary linkages that are most worthy of investigation? We need an exercise in collective triage.
Your point about different levels of organization is a good one. But you then recognize that sometimes interactions between phenomena at different levels are important. Is there some sort of guideline we could generate to distinguish the important cases from the ones we can safely ignore, or do we need to exercise our collective judgment case by case?
Thanks again, Rick
Good question, Rick: can we recognize or even predict when inter-level interactions are creating ‘unknown unknowns’ compared to those interactions that do not? In many cases, it is difficult in itself to make sure that we are not conflating levels. For example, obviously quantum phenomena occur in neurons that are relevant for cognitive functions. However, recognizing the impact of these quantum phenomena on individual neurons is not the same as recognizing that these phenomena have an impact on the cognitive function to which such neurons (alongside with millions other neurons) contribute. So in a way quantum phenomena do ‘affect cognition’, namely they affect constitutive elements of cognition yet they do generally not affect the behavioral or cognitive outcomes of these elements. Until further research does show that they slightly modify these outcomes, of course… I’m not sure this really helps, although it reminds us of remaining open-minded about potential new explanatory relations between ‘everything’ and ‘almost everything else’…
This idea of unknown unknowns makes me think of the problem on how best to disrupt racist actions amongst law enforcement (a “wicked problem” per PHILOSOPHY). The recent PSYCHOLOGY research demonstrating that implicit bias training changes attitudes, but not necessarily action, poses a challenge to social scientists (https://www.npr.org/2020/09/10/909380525/nypd-study-implicit-bias-training-changes-minds-not-necessarily-behavior). SOCIOLOGY/critical criminology research calls for systemic action, at a minimum, this means tracking racist actions/outcomes and imposing oversight to disrupt, and ideally this means abolition movement. Alternatively, clinical SWK [social work] research offers us theories of racialized trauma and embodied racism, calling for individual-level approaches to disrupting racism (e.g. See Menakam’s My Grandmother’s Hands: Racialized Trauma and the Pathway to Mending Our Hearts and Bodies which provides the theory and the call to action for law enforcement to do this personal work). This work can explain why the implicit bias training falls short. And it can explain, in part, why systemic approaches are so stubbornly resisted by so many who profess a desire to see real change.
Thanks Jessica! Is it your sense that some combination of these strategies would work better than each does on their own? Or is there some piece of the puzzle that they all are missing?
Thanks, Rick, I like this a lot. It is closely related to the system-of-systems approach, which has been around for a long time, and which we are trying to apply to complex socioenvironmental systems, as described in these papers:
Little et al 2019 A tiered, system-of-systems modeling framework for resolving complex socio-environmental policy issues
Iwanaga et al 2021 Socio-technical scales in socio-environmental modeling – managing a system-of-systems modeling approach
All the best
Thanks John! I will take a look. I like the phrase system-of-systems a lot, Rick
Rick–a follow up comment: you may enjoy this article, originally published in Science https://www.researchgate.net/publication/316286606_Ecosystem_management_as_a_wicked_problem
I happened to be reading it for a piece on business ecosystems and it dovetails nicely, I would suggest, with your work
Thanks Jim for your comments and advice! They are especially welcome as my current book project, Making Sense of the Future, is an exercise in Foresight (informed by my experiences in world history and interdisciplinary studies).
I agree with you that we need to understand how different trends potentially interact, and then carefully identify and experiment with ways of pushing systems in desired directions.
Complexity is a challenge, but we are better able to cope with it when we face up to it.
Thanks again, Rick
Rick–that is a creative framework–unknown unknown–and it seems to highlight our difficulty trying to discern effects in light of complexity and emergence. The discipline of foresight uses morphological analysis to consider the effects of clashing trends in different areas (e.g., social, technical, environmental, economic, political and ethical) and one challenge there is knowing where to apply constraints in the framework. Doing it prematurely can limit insights while applying too broadly makes it difficult to pull out meaningful insights. Thanks for the thoughtful article. Jim