Learning with a place

01 Learning from a place

In the autumn of 2025, volunteers stood along Bowker Creek in Oak Bay and watched for chum salmon.2 Thirty-six thousand eggs had been incubated by hand in boxes in the creek in 2023.3 Chum are reported to return three to five years after hatching.4 People went down to the water and looked.

None came.1

That is a fact about the world which no amount of planning produced and no database could have supplied. It arrived the way ecological knowledge usually arrives: someone was standing in a particular place at a particular time of year, paying attention, and the place did not do what was hoped. The group gives two reasons — chum returns were generally low that year across Salish Sea rivers and streams, and its own first incubation boxes turned out not to suit the creek.5 The three-to-five-year window means that batch has not finished being answered for; the collection puts the next possible returns at 2026 and 2028, and marks that as its own reading of the window rather than something any source states.6 The record is open, and open is what it is.

This is the first and least fashionable kind of learning: the place tells you something, and it is often not what you wanted. It has a method, a date and a limit attached — who looked, where, when, and what would have counted as seeing a fish. Strip those off and you do not have a stronger fact, you have a weaker one wearing a confident expression.

02 Learning with the people in it

Bowker Creek drains a watershed of either 1,028 hectares7 or 1,850,8 depending on which document you open — the plan's own figure and the Friends of Bowker Creek's, and this collection publishes both rather than picking one. The creek runs from the University of Victoria through Saanich and Victoria and into Oak Bay,9 9.4 kilometres of it, of which 3.4 remain open channel and the rest is pipe and culvert.10 Three municipalities, no shared council.

Those three councils endorsed a hundred-year plan for it between September 2011 and July 2012.11 School District 61 added a fourth endorsement in March 201812 — five years and eight months after the last council, six and a half years after the first, on the collection's own arithmetic over those dates. Consent to a long plan arrived in instalments.

The plan numbers seventeen reaches of the creek13 and thirty-three watershed actions, each naming a responsible organisation and a potential funding source — with the plan's own warning that funding sources will shift and the column will need re-evaluating.14 It does not name an end year: its own text puts the horizon at as long as 50–100 years from now,15 which is a range, and no source in this collection converts it into a date.

It also says, in both appendices, that it cannot mandate any group to carry out the actions.16

The people implementing the plan named the problem themselves, out loud, a decade in. What one of them concluded holds a hundred-year promise together is a part-time coordinator, cost-shared between three municipalities;18 and the steering committee asks prospective public members for a two-year commitment.19 A two-year ask inside a hundred-year plan is not a contradiction anyone failed to notice. It is what a voluntary arrangement between separate authorities can actually offer.

So the interesting question about Bowker Creek is not what should be done to the creek. Thirty-three actions already answer that. It is who keeps going, and how, when nothing compels them — and that question is not ecological. It is about relationships, obligations, money, attention, fatigue and who is still in the room in 2040.

Which is why mapping a bioregion cannot only mean mapping a bioregion. A map of the watershed does not show the two-year commitment, the cost-share, the four endorsements gathered in instalments, or the fact that a 2021 assessment — which reaches this collection quoted inside a 2025 brief rather than read directly — names two individuals as the people the whole thing runs through.20 Those are the load-bearing facts. A mapping practice that captures the creek and not the arrangement has mapped the easy half.

The workshops this project comes out of try to map the other half too: the questions people actually hold, the disagreements they have not resolved, what they want, what they are afraid of, what they can offer, what they need, and what they might be willing to commit to. Those are not softer versions of geographic data. They are a different kind of thing, they change when you ask about them, and several of them are nobody's to publish.

03 Learning as a place

The third preposition is the strange one, and it is worth being careful with it.

Bowker Creek was buried. The Friends of Bowker Creek's director puts the timescale as a century each way: it took the whole of the twentieth century to bring the creek down, and will take the twenty-first to bring it back.21 An open watercourse became culvert and pipe under streets and schoolyards, leaving 3.4 of its 9.4 kilometres open to the sky.10 Burying it did not only remove a creek; it removed the possibility of noticing one. You cannot watch for salmon in a pipe. You cannot have an opinion about a creek your children have never seen.

That an infrastructure decision propagates into what a place is subsequently able to perceive about itself is our argument, not a finding, and nothing here measures it. The nearest thing to a measurement in the collection is smaller and about the plan rather than the creek: the 2023 update to the plan's own corridor appendix records that a significant number of residents encountered during the 2022 field survey had never heard of the Blueprint or its restoration objectives.22 A plan that works by asking landowners to volunteer depends on landowners knowing it exists.

Daylighting reverses the burial, slowly, reach by reach. The plan's Action 7 tells partners to pursue daylighting as part of changes to land use or when hydraulic structures are being replaced23 — restoration attached to somebody else's timing rather than to a restoration budget. At Oak Bay High School it worked: ten years on, the CRD reports that reach restored through a collaboration between the school district, Oak Bay and the Initiative.24 What we think follows from that — people begin to see water where there was pavement, which changes what they ask, which changes what the next council meeting is about — is a claim of ours with nothing behind it in this collection.

04 Learning between places

Nobody in Victoria needs to be told that volunteers get tired. What a wider commons can offer is narrower and more useful than advice: an account of what other places tried, and under what conditions it held.

Seoul's Cheonggyecheon is the case people reach for first, and what this collection holds about it is the awkward half. Water is only naturally present in the channel during the summer rainy season, so it is kept flowing by pumped and treated water from the Han River and subway pump stations — on the account of a 2011 case-study brief, which is the only figure we hold.26 A later retrospective found that the public sector could not keep management consistent, because the project became a matter of political upheaval at every four-yearly mayoral election.27

Zürich, in the words of the article this collection quotes, “began with pilot projects in the 1980s and restored numerous smaller stream segments over several decades”, and “linked the projects to sewer management, infrastructure renewal, ecological restoration and neighbourhood planning”28 — many small segments riding cycles that were going to happen anyway, rather than one flagship. Still Creek, across Vancouver and Burnaby, went the same way: “site-specific projects involving habitat restoration, rainwater management and public access, including a developer-funded daylighting project completed in 2009”, over more than two decades, “rather than one defining project”.29

Seattle's Green Seattle Partnership marked twenty years in 2025 and published three counts: 494 Forest Stewards trained since 2005, 171 described as active, 71 with ten or more years of service.30 The page defines neither trained nor active, so those are three counts and not a retention rate, and this essay does not turn them into one. Twenty years is a fifth of a century — but Bowker's own horizon is a range, fifty to a hundred years,15 so the comparison is looser than the round number sounds. And after twenty years that partnership is moving to a three-year action plan rather than another twenty-year one.31

One pattern is worth carrying home, and it is our reading across those records rather than something any of them states: the arrangements that held named somebody to be responsible and paid for the upkeep, instead of leaving maintenance to whoever is still enthusiastic. The only fully worked version of it in this collection is a United States rule for permitted wetland impacts, and the modal verbs are the point — naming the responsible party is mandatory, while costing the work and naming a funding mechanism is only recommended.32 The land-trust sector that does the arithmetic behind such endowments says plainly that it is a mistake to plan on volunteers handling stewardship responsibilities in perpetuity.33 Toronto is the counter-case the collection carries: policy support for restoring buried streams which, in the same article's words, “fails to identify what sites should receive priority, who should assess them or how a project would be funded and maintained”.34

And the research on volunteer attrition adds an uncomfortable corollary — decline is structurally invisible until it is acute, because volunteers rarely quit formally, they simply stop showing up.35 Nothing in this collection establishes what a typical volunteer restoration manages before it stops, and this essay is not going to supply a number for it.

That asymmetry is the point. Local people know their place. A translocal commons can know what has been attempted elsewhere, including what failed — a reading of an idea the other collection on this site records as cosmo-localism, where heavy things stay local and light things travel36 — and the second kind of knowledge is only useful if it stays visibly the second kind. When you ask a question here and the answer draws on both, the answer names which collection each passage came from, because a municipal plan for a specific watershed and a case study from another continent are different sorts of evidence and should never be quietly averaged.

05 The cycle, and why it is a spiral

Put the four prepositions in order and they make a loop: someone notices something, people interpret it differently, a decision gets made and resourced, something happens, and the place answers. Then the loop runs again — but not from where it started, because the first pass changed what the second pass is able to see.

One turn, written out

Here is what a single turn looks like when the stages are kept distinct. The first two rows carry the same numbered marks as the essay above. The third is our reading across the translocal records. The last two are not evidenced at all — and are marked accordingly.

The two greyed rows are the honest part. It is easy to draw a complete loop and much harder to admit that yours stops after four stages because the fifth requires a person to decide something and nobody has. A diagram that completes itself with imagined participation is worse than no diagram.

06 What a commons like this can hold

Not the learning. The learning is in the volunteers who know which bend floods, in a council's habit of endorsing things, in the creek's own hydrology. What software can hold is narrower: a record that keeps its provenance, so a claim can be argued with; a way to notice what another place tried; and a route back to the people whose knowledge it came from.

Which is why the two working demonstrations on this site are the ones they are. The Atlas shows records that name their own sources and a scope line that names its own losses — you can see what is missing before you see what is there. The exchange shows a note travelling between two independently run commons where the receiving one decides for itself whether to accept it, because the alternative — a single index everyone must join — is how you get one community's vocabulary quietly overwriting another's.

And the published collection here is small on purpose: thirty-four records,37 each traceable to a page you can open. A much larger working collection exists and is not published, because its provenance is private meeting notes and no consent covers putting those on the web.38 The gap between those two numbers is not an embarrassment to be engineered away. It is the actual condition of every knowledge commons that has real relationships behind it.

— References

Every small number in the essay links to a passage here. Every passage links back to the sentence it supports, and out to the document it was read from.

A check runs here when JavaScript is on: it re-reads each quoted passage from the collection the gateway is serving and reports any that no longer match. It has not run.