In Brief
The academy has one wall left standing once the internal ones come down, and it runs around its own tradition of knowing. Breaking the barrier between engineering and ecology still leaves an institution working inside a single epistemology.
Stewardship read as a value produces respectful language. Read as a method it produces testable causal claims which can directly influence what a research programme actually does.
The question of who validates a finding has been answered by Australian Indigenous scholars for three decades. The gap is institutional adoption, not intellectual availability.
Martu understood how fire, hunting and desert fauna interact long before conservation ecology tested the proposition and found it held. That is an academic result, not a cultural courtesy.
Kachana and West Arnhem Land show what turns on whether a research partnership forms at all. At West Arnhem one did, years before any market for the outputs existed, and one was later built on what it produced. At Kachana none did, and four decades of work have never been independently verified.
Indigenous Data Sovereignty principles and the AIATSIS Code are already funding conditions. Institutions carry the cost and extract none of the methodological benefit.
The previous essay closed with a window on the hardest of the walls for a university to break down. In order to observe a whole system, an institution has to be willing to look past its own way of seeing, and to treat another and older tradition of managing complexity as a partner rather than a subject to study or an ethical consideration.
That is easy to write about and hard to implement not only because it relates to subjective values and judgement calls, but also to methodological processes and ways of working. An institution that sets out to walk alongside a knowledge system it has little to no capacity for or contextual background with has to answer important questions about validity. Who decides whether a finding is useful, where it isn’t applicable and for what purpose?
Australian Indigenous scholars have been responding to these challenges for three decades. Martin Nakata, a Torres Strait Islander scholar, describes the cultural interface as the contested space where two knowledge systems meet and neither reduces to the other (Nakata, 2007). Aileen Moreton-Robinson, a Goenpul scholar of the Quandamooka nation, has traced how possession works as the organising logic of settler institutions, including the standards those institutions use to decide what counts (Moreton-Robinson, 2015). Poelina et al. (2019) document what changes when Traditional Owners hold the governance of a catchment rather than being consulted about it.
The answers are within reach but what is missing is any signal that institutions have picked them up or are even curious about discovering them.
If we return to the archetypal stream this series has been floating along since the initial instalment and look back to Who Owns the River?, that essay argued that the Aboriginal Waterways Assessment (AWA) turns cultural condition from a consultation footnote into a decision-relevant signal. Without this, an agency allocating water is working from an incomplete picture of the system it is engaging with. The setting there was a choice already on the table, with an owner, a cadence and a file to be closed.
The scenario explored here sits closer to the source of information flows. Before evidence of that kind gets to the point of redirecting allowances, somebody has to generate it, validate it and stand behind it. The AWA exists because Traditional Owner organisations, a basin authority and researchers built a robust assessment process together over several years. That construction work is part of the university’s core business, and is what remains to be examined in the transition from inertia to immersion.
Handled badly this work looks like extraction and appropriation with better manners but done well it becomes a credible, repeatable and verifiable method, one that the academy is uniquely placed to run and almost entirely failing to.
The Wicked Problem of a Single Tradition
Even for people working within these institutions who are not indifferent to a more collaborative approach the constraint is an architecture which turns a question into admin before anyone can treat it as inquiry.
Inertia here is a form of structural blindness with Indigenous knowledge sitting in one faculty, ecology in another, accounting in a third, and the ethics of working across them sits with a committee that meets monthly and issues approvals. Each part functions individually exactly as it should but there is no structure that asks them to combine around a research question, which is the same fragmentation the previous essay described, now running across traditions rather than across departments.
As it stands, compliance procedures make this worse rather than better. An ethics approval is a gate to pass and once it has been satisfactorily agreed with, the file closes. Nothing in that process asks what the partner knows, whether their account of the system is testable, or what would happen to the research design if it turned out to be right. The institution has installed an obligation and left the method undiscovered inside the mechanism.
The problem does go deeper than just having an unhelpful box to check though. Anyone who actively wants to work this way runs out of forms to fill or avenues for support to explore. There is no funding category for paying a knowledge holder who is a collaborator rather than a subject, no template for a thesis where part of the argument is carried by something other than text, and no convention for crediting a contribution that does not resolve into authorship as the institution defines it. What meets them is not flat-out refusal but an absence of shape to work within, which is far harder to argue with, because at least rejection comes with an appeals process.
There is also a seemingly intractable circularity built into the ethics process. Co-design means the partner helps set the research question, so the question does not exist until the conversation has happened. An ethics application requires that question stated in advance, and approval must be in hand before the conversation can begin. Researchers working this way describe several rounds with a committee to resolve a loop that the form is unable to hold. It’s unlikely that this trap was consciously designed but is rather something that falls out of a procedure which assumes the question at hand has been fully formed, which is precisely the pre-condition that this essay is asking institutions to re-consider.
That is the inertia. Immersion is the other state, and in this setting it means treating another tradition as a source of evidence with its own standards rather than as a population to be consulted.
Stewardship as a Method, Not a Value
The critical distinction here is between respecting a tradition and using one.
Respect is a posture. It asks for humility, acknowledgement and care in language. All of that is worth having, but it still means that a research programme runs exactly the same way it always has next week regardless.
Method is different. Enqvist et al. (2018) treat stewardship as a boundary object with three working dimensions.
Care supplies the ethical commitment and the attachment to place.
Knowledge supplies the competencies that inform action.
Agency provides the capacity to act at all.
Chapin et al. (2011) push further, framing custodianship as a science for action rather than a set of values held about nature. Both emphasise that a long-standing practice is not a belief but a decision rule with a causal claim inside it, and causal claims can be tested.
Two Eyes, Held Apart
Mi’kmaw Elder Albert Marshall offers the principle of Two-Eyed Seeing, or Etuaptmumk. Learn to see from one eye with the strengths of Indigenous ways of knowing, from the other with the strengths of Western science, and use both together (Bartlett, Marshall & Marshall, 2012).
The formulation is precise and this is key to effective implementation. Two-Eyed Seeing holds both perspectives intact without playing favourites. Each is kept whole and separate, and the practitioner moves between them as the situation requires. Marshall’s own framing is that neither eye is asked to become the other.
The importance of this approach is practical rather than philosophical. A blended framework has one validation standard, and in almost every case it turns out to be the Western one, because that is the tradition doing the blending. Holding the perspectives apart preserves two independent tests. A finding that survives both is stronger than a finding that only survives one or the other.
“A blended framework has one validation standard which generally turns out to be the Western one, because it’s the one doing the blending.”
A prediction the discipline could not hear
In Australia’s Western Desert, Martu people hunt with fire. Small patches are burnt in the cool season to flush game, and the result over time is a carefully constructed, fine-grained mosaic of vegetation at different stages of regrowth.
Rhys Jones named the practice fire-stick farming in 1969, so the proposition that Aboriginal burning shaped Australian ecosystems has been available to the discipline for over half a century (Jones, 1969). Availability and uptake are different things. Arid zone conservation continued to work from the assumption that reducing human pressure would benefit native fauna, and that hunting in particular would suppress prey numbers.
Bliege Bird et al. (2012, 2013) found the opposite. Sand monitor populations are higher where Martu hunting is most intense, because patch burning rescales habitat diversity at spatial scales the animals actually use. The same burning buffers the fire-size variability that climate alone would produce. Where people had left, the mosaic collapsed into large uniform burns and the fauna went with it.
Martu had held this understanding for a long time and had stated it in terms the Western discipline could not initially hear. The account attributed the mid-century wave of small mammal extinctions to the removal of people from Country rather than to their presence. That prediction pointed the right way. The mechanisms are now understood as interacting, with invasive predators and pastoral pressure compounding the loss of the fine-grained mosaic, but the direction of the claim held, and it was stated long before the ecological literature caught up.
There is an epistemic consequence here because Martu held the better causal model while Western ecology eventually arrived at the same position, much later, by a different route and having first assumed the reverse. The real value of this delayed recognition is the understanding itself rather than in its usefulness to a conservation objective someone else set. Nyalangka Taylor appears as a named author on the resulting paper, which means there is a community voice inside the citation rather than beside it but even that level of participation and recognition has limitations. Authorship compresses, and what gets discarded in a compression is decided by the receiving system rather than by the sender. Knowledge held in common goes through the same narrowing on the way to a list of names. The clearest available example of this method working still shows an institution shaping knowledge to fit a format it was comfortable with.
This belongs in the literature rather than in the ethics file. Working across two knowledge systems overturned a standing assumption in the field. Not extra context but a correction to the record.
As great as this outcome was, what happened in the Western Desert is not Two-Eyed Seeing in full. Martu knowledge supplied the hypothesis. Western ecology tested it, using its own instruments, and published in its own journals. A single validation standard was operating and it belonged to the discipline. That is hypothesis generation across a boundary, which is worth a lot while still falling well short of the ceiling.
Nakata’s cultural interface describes what the fuller version requires, which is a space where both systems keep their own tests and the friction between them is treated as data rather than as a problem to be resolved. This is where the opportunity for real change sits.
Two Landscapes, One Absence
What is at stake is which questions get asked, for how long and by whom. Two remote parts of Australia, each managed for decades by people who knew what they were doing, set out the problem clearly.
Kachana Station occupies 77,500 hectares of the East Kimberley, in the upper catchments of five river systems, reachable by aircraft or a four-day walk. The Henggeler family arrived in 1986 and settled permanently in 1991. What they found was a landscape shedding itself, bare ground, eroding gullies, water running brown. Almost four decades later the ground is covered, springs have returned, gullies are filling and the creeks run clear. The family funded this themselves, without institutional support, over a period longer than most research careers.
The regenerative methods they use draw on Savory, Ingham and Jones, and on holistic management, which remains contested in rangeland science. That dispute is part of the point rather than an awkward footnote to it. Four decades of monitored data from a working property would have gone some way towards settling the question, and only a fraction has been systematically collected. The family have documented the change for decades, through photo points, journals and close observation. What that record lacks is not diligence but the scaffolding that would make it admissible, a baseline, an agreed method and verification by someone without a stake, and so the question stays open on both sides.
Kachana has since become a lighthouse project for Earthly Treasure, the not-for-profit arm of The Landbanking Group, where I hold a role as APAC Partner, and some of the reasoning that follows is drawn from that involvement. Declaring that connection is not merely a formality because stating where an interest comes from is what lets a reader locate the blind spots and the likely direction of any confirmation bias, which makes it part of the method rather than a pause in it.
The family earn what a working station earns, cattle sales that meet the conditions of a pastoral lease and workshops for those who make the journey. For the ecological function they have restored, the return has been nothing. Not less than expected. None.
The structural reason for this is that environmental markets pay for measurable change against a counterfactual, which is a modelled account of what the site would have done had nobody intervened. Carbon schemes pay for emissions avoided or sequestered against an approved method and baseline, and Kachana has already made the change it would need to sell. Australia’s Nature Repair Market is a worthy attempt to change this although there are open questions around the demand signals which will make this a functioning scheme. The Nature Repair Act contemplates protection alongside improvement, and methods for permanent protection, remnant vegetation and the rangelands are all in development.
So the most honest position is not that no instrument could ever see this country. It is that the instruments capable of actually capturing its qualities are still being built, and that every one of them will ask the same question when it arrives. What is the verified condition of this place, measured how, against what baseline, and over what period. Four decades of work without a monitoring programme cannot answer that.
There is a second absence, and for the academy it is the more serious one. Four decades of continuous change across tens of thousands of hectares is an unusually long record of rangeland recovery in northern Australia. Researchers have visited, papers have been written, a documentary has been made. What never formed was a sustained programme of coordinated sampling with a baseline, an agreed method and a commitment to keep measuring.
Several hundred kilometres east, the same absence was filled, and everything downstream followed from that. West Arnhem is the counter-case, and the difference is that a university turned up and stayed long enough to get a result that mattered. The fire abatement project there began in 2006 alongside Traditional Owners, Indigenous ranger groups, the Northern Land Council, the Territory government and bushfire researchers from Charles Darwin University. They were partners from the start rather than analysts brought in afterwards. The project became the model on which Australia’s approved savanna burning methodology was built, and Aboriginal-owned entities have since earned millions of carbon credits, with revenue directed to community priorities including rock art knowledge recording, ecological monitoring and women’s ranger programs.
That is a real achievement which should be interpreted carefully, because it is possible to get the causation backwards. The market did not create value at West Arnhem. The knowledge produced a method, the method produced an instrument, and the instrument then bought the single attribute it was capable of pricing, carbon. Fire practice does more for Country than reduce emissions which is something the practitioners can readily attest to but none of that entered the transaction.
“An instrument can only pay for what it can measure, so what gets valued depends on what has been recorded.”
Set these two landscapes beside each other and a pattern emerges. Different traditions, different tenure, and the same constraint. An instrument can only pay for what it can measure, so what gets valued depends on what has been recorded. The evidence comes before the instrument rather than after it, which is why the schemes now being built to price condition held will still ask Kachana a question it cannot answer.
The dates matter here. The partnership formed in 2006 under a bilateral contract with a gas company. The Carbon Farming Initiative legislation did not exist until 2011. Registration as an eligible offsets project came in 2014, and the savanna burning methodology determination is dated 2015. Nearly a decade of work came first. No market arrived to commission it. The account existed, and the instrument was designed to attach to it.
Kachana has no equivalent. Four decades of management and no verified account of what changed, which is why nobody can price it and why no sustained programme has ever been built on it. Those are not two separate failures. A grant panel can fund exactly that gap, and sometimes does, but somebody has to be positioned to make the case and to sustain the relationship once funded. At West Arnhem somebody was. At Kachana nobody has been, and remoteness, private pastoral tenure and a contested management approach have each made the case harder to mount.
Which returns the essay to where it started. An account of a system is not a neutral object. Somebody decides what counts as verified, to what standard, and how often the question gets reopened. Who decides whether a finding holds.
What the Balance Sheet Decides
The accounts make the same call using the same logic.
Financial reporting gates recognition on benefits that accrue to the reporting entity, so the test turns on where the return is channelled rather than on how well the condition is measured. At Kachana four decades of work appear as four decades of expense and no asset, because the water quality, the moderated flood flows and the reduced fire exposure all reach someone else. IPSAS 51, issued in January 2026 by the International Public Sector Accounting Standards Board and effective from 2028, moves part of that line for public sector entities. Yet the same standard treats the labour of keeping a resource in good order as an expense unless it measurably increases what the resource can deliver, and its own worked example of such expense is rangers employed to protect the site. Buying more country builds an asset. Looking after it remains a cost, and the people doing the work of maintaining system integrity become a line item to be minimised.
Every standard is published alongside a Basis for Conclusions, which sets out what the Board was told during consultation and why it decided as it did. Indigenous respondents advised that, given their traditions and practices, it may be difficult for any entity to demonstrate control over natural resources. Control is the concept that gates recognition. People whose custodial relationships predate the accounting frameworks by millennia were telling a standards board that its central test does not describe how they hold Country.
Moreton-Robinson’s account of possession reads as direct commentary on that exchange. A standard that gates recognition on control is not making a neutral technical choice. It is encoding one particular relationship to land as the only relationship the ledger is built to see and then treating everything outside it as a measurement problem.
Where Authority Actually Sits
Consultation, engagement and authority are different things, and most partnerships fail in the space between them.
Consultation asks. Engagement considers. Authority decides. A research partnership can consult extensively, engage with multiple stakeholders and still hold every decision that matters about what gets measured, what counts as validation, who publishes, what incentives are pursued and where the benefit is directed.
The 2019 fish deaths in the lower Darling show what that costs. Two independent reviews were commissioned. Neither included an Indigenous knowledge holder, because Barkandji people were not treated as experts on a river their families have read across thousands of generations. Australia’s State of the Environment report records this, and records alongside it that Indigenous knowledge continues to hold lesser status for influencing policy and legislation than non-Indigenous knowledge (Cresswell, Janke & Johnston, 2021). Consultation and engagement were available throughout. Authority never was.
Australia has an Indigenous-authored answer to this. The Indigenous Data Sovereignty principles developed and endorsed at the 2018 Canberra summit convened by Maiam nayri Wingara set out the terms on which data about Indigenous people, resources and Country should be governed. Internationally the CARE principles sit alongside FAIR, adding collective benefit, authority to control, responsibility and ethics to the technical criteria that govern data management (Carroll et al., 2020).
The institutional point is sharper than most entities might realise. Compliance with the AIATSIS Code of Ethics is already a condition of funding from AIATSIS itself, the Australian Research Council and the National Health and Medical Research Council. Universities carry this cost now. They carry it almost universally as an ethics approval burden, discharged early and filed, and they extract no methodological benefit from it whatsoever.
The gap between a compliance obligation and a research method is where most of the available value is sitting.
Two Tests at the Boundary
The full essay meets five objections directly, sets out three diagnostic lenses and offers four tests for any knowledge partnership. Two of the tests are worth stating here.
Name the claim, not the value. Ask what causal assertion the practice carries. When a freshwater crayfish walks backwards out of a waterhole, the instruction attached is to stay out of that water. Underneath the instruction sits a claim about the chemistry, most likely dissolved oxygen or pH, made by an animal that reads the system faster than a sampling round can. Indicators of this kind have been catalogued in large numbers on Country in western New South Wales, each one a prediction with conditions attached. Write both down and treat them the way you would treat any hypothesis arriving with thousands of years of observation behind it.
Test where authority sits. Can the partner refuse, and does that refusal stand? Can the partner hold the benefit rather than receive an acknowledgement? If either answer is no, the arrangement is not a partnership and the label will not fix it.
The Evidence No One Else Can Generate
The academy has a specific asset here and it is not moral standing.
Having a longitudinal series of metrics and indicators is not necessarily where the value is here, because Australia already does data reasonably well. The Terrestrial Ecosystem Research Network holds monitoring plots across the continent, the state herbaria hold collections going back generations, and agency programmes run for decades. What none of that infrastructure holds is a relationship with the people who can read a place directly. Records survive staff turnover. Trust does not, and trust is what determines whether anyone still wants to be working with an institution in twenty years.
A university has the potential to hold relationships across many decades. A consultancy cannot, because the engagement ends. A firm cannot, because ownership changes. An agency struggles, because programmes are rebuilt with governments. The multi-generational evidence base that would make ecological condition legible needs an institution with a long memory and a reason to keep it, and there are very few of those.
West of Kachana, Traditional Owners of the Martuwarra have built a council to govern their own river and are asking for a statutory authority to hold it, which is the same case being made from the opposite direction (Poelina et al., 2019). At Kachana five river systems run clear out of country that was bare four decades ago, and no ledger anywhere records it.
Three catchments, three different constraints. The basin was over-allocated by licence and by politics long before anyone asked what counted as evidence. The Martuwarra Council already holds both the understanding and the governance structure, and what it lacks is statutory standing. Kachana lacks the record itself. What runs through all three is something more specific than a common cause. In each case the people who know most about how the system behaves have no route by which that knowledge becomes something a decision has to answer to. The academy is not the whole answer to any of the three but is the institution best placed to build that route. Doing so is a design choice made in ordinary places, in the wording of an agreement and the terms of an approval.
Inertia is a signed ethics approval in a closed file. Immersion is a research design that changed because other perspectives were not only sought after but were also found to be more instructive.
Read the full essay at emerdigm.com, including the three diagnostic lenses, all four tests, the five objections and the complete reference list.
Next in the series we move from the campus to the board room in “Beyond ROI: What the Balance Sheet Cannot Hold”, where the same analysis is applied to the accounting frameworks themselves, and to the question of what happens to value that no market prices and no standard records.
A note on this essay. It draws on published research, including work co-authored with community members, and on governance principles authored by Indigenous organisations. It has not been reviewed by the communities whose practices it discusses. Corrections are welcome and will be published.
Key sources
Bartlett, C., Marshall, M., & Marshall, A. (2012). Two-Eyed Seeing and other lessons learned within a co-learning journey. Journal of Environmental Studies and Sciences, 2(4), 331–340.
Bliege Bird, R., Taylor, N., Codding, B. F., & Bird, D. W. (2013). Niche construction and Dreaming logic: aboriginal patch mosaic burning and varanid lizards in Australia. Proceedings of the Royal Society B, 280(1772).
Cresswell, I. D., Janke, T., & Johnston, E. L. (2021). Australia State of the Environment 2021: Overview. Canberra.
Moreton-Robinson, A. (2015). The White Possessive: Property, Power, and Indigenous Sovereignty. Minneapolis: University of Minnesota Press.
Nakata, M. (2007). The cultural interface. The Australian Journal of Indigenous Education, 36(S1), 7–14.
Poelina, A., Taylor, K. S., & Perdrisat, I. (2019). Martuwarra Fitzroy River Council: an Indigenous cultural approach to collaborative water governance. Australasian Journal of Environmental Management, 26(3), 236–254.
Full reference list in the complete essay.