LL2-17 — Ch17 Ecosystems and managing the dynamics of change#

Late lessons from early warnings: science, precaution, innovation (EEA Report No 1/2013), Part “Some emerging issues”. Report pages 407–428 (PDF pages 409–430). Main text pp. 407–424; Box 17.2 (definitions) pp. 424–425; references pp. 425–428.

Read in full from the text extract (all 22 page markers, through PDF 430 / report p. 428). Pages 407, 410, 412–418, 421, 423 and 424 were also checked visually in the PDF: the abstract, Box 17.1, Panel 17.1, Tables 17.1–17.3, Figure 17.1 and Panel 17.2. The extract is clean. The one layout trap is that Panel 17.1 (pp. 412–414) interrupts the main text. The main-text list of assessment flaws begins on p. 411 (“First…”) and resumes below the panel on p. 414 (“Second…”, “Thirdly…”). The notes below keep the two streams apart.

Authors and standpoint#

Chapter authors: Jacqueline McGlade and Sybille van den Hove (p. 407). The chapter gives no affiliations, but the report itself establishes where the authors stand:

• Jacqueline McGlade signs the report’s Preface as “Professor Jacqueline McGlade, Executive Director” of the EEA (p. 8). This chapter is therefore written by the head of the agency that published the report. It speaks with the institution’s voice and is not an external case study. Panel 17.1 adds that in 1980 she was “a newly recruited scientist in the Department of Fisheries and Oceans (DFO), in the Canadian federal government”, assessing pollock stocks (p. 412). She is thus an insider-witness to the DFO assessment culture she criticises. The chapter leans heavily on her own earlier work: McGlade 1999, 2001, 2002, 2003 and 2009, Hendry & McGlade 1995, Hogg, Huberman & McGlade 1989, and Dublin, Sinclair & McGlade 1990 (Stokes, McGlade & Law 1993 is in the reference list but not cited in the text). The core analytical device, “diversity‑oriented” versus “variable‑oriented” approaches, is credited to McGlade 2003 (pp. 408, 424–425).
• Sybille van den Hove is on the report’s editorial team (Acknowledgements, p. 5). Her work on “consensus and compromise” in participatory processes (van den Hove, 2006) is cited once, in the future-prospects section (p. 423).

Panels and boxes, with authorship:

No panel in this chapter dissents from the chapter’s line, and there is no industry or regulator response. The only “other side” voices are quoted as evidence against themselves: Fisheries Minister John Crosbie’s three-factor explanation of the collapse (p. 411), the DFO Director General in Fo’c’sle (p. 414), the Assistant Deputy Minister’s “official line” (Parsons 1993; p. 413) and the Canadian Minister’s legal argument in the Nunavut case (p. 420). The one official quoted with approval is Norway’s Fisheries Minister, and only in reported speech (p. 411).

Evident stance. The chapter is conceptual and essayistic. It is not a standard single-hazard “late lesson” history. It argues for four things:

1. Close, long-term observation of nature.
2. Configurational, “diversity‑oriented” analysis in place of reductionist, equilibrium, “variable‑oriented” resource models.
3. Wider knowledge inputs, including lay, local and indigenous knowledge.
4. Resilience, panarchy and institutional-fit thinking.

It is strongly critical of the equilibrium models used in fisheries (MSY, F0.1) and of political decisions that override scientific warnings. It is pro-precaution, but its main target is epistemic and institutional: how knowledge is produced, filtered and acted on. It is less about precaution as a decision rule. It sits under the report’s “Some emerging issues” part (running header), and is more a reflection on ecosystems generally than a report on one emerging technology.

Section-by-section notes#

Abstract (p. 407)#

• Framing. Carson’s Silent Spring and Annie Dillard’s Pilgrim at Tinker Creek (1974). Close observation of nature holds “the seeds of surprise — surprise of the unusual and surprise as a portent of future change”. “Our systemic failure to anticipate such surprises forms the core of this chapter” (p. 407).
• What the abstract promises. Case studies “from fisheries, forests, savannah and aquatic systems” showing “how early warnings about changes in these natural systems emerged but were not used”. Only fisheries is treated at length. The other systems appear as rows in Table 17.2.
• Diagnosis. Knowledge divided into “political, disciplinary and geographic silos” produces “recurring nightmares” (attributed to Yaffee 1997 at p. 415): short-term interests beat long-term vision, competition replaces co-operation, and values, authority and information fragment. “Lack of institutional fit” brings “unexpected surprises, excessive rent seeking and high transaction costs” (p. 407).
• Method. “Counterfactual thinking”, built around planetary boundaries, tipping points, panarchy and resilience, is offered as “an analytical lens” on why warnings “were not seen”. The chapter ends by saying ecosystems will be “even more at risk”. We will need to watch “the dynamics of both nature and institutions ever more closely” (p. 407).

17.1 Introduction (p. 408)#

• Epigraph. The section opens with Carson (1962) on the “deceptively easy” superhighway that ends in disaster and the road “less travelled by” (p. 408).
• Examples of close observation: Dillard, and Charles Keeling (“dubbed the father of climate change science”).
• History as insight. Collingwood’s woodsman who sees “a tiger in that grass”. Ecological memory: a tree’s age and a forest site’s history shape present vulnerability (Hendry & McGlade 1995) (p. 408).
• A general theory is emerging. Comparative studies are building a general theory of socio-ecological change (Walker et al. 2006). These studies show “just how limited many more classical scientific approaches really are for dealing with real‑world situations”: simple causal models, controlled experiments and statistical sampling. “The continued dependence of many resource managers on these more traditional approaches has contributed to the catastrophic decline of some resources” (p. 408).
• Aim. To explain “why the evidence was sometimes flawed, how and why scientific advice influenced or was ignored” (p. 408).

17.2 Data, knowledge and counterfactual thinking; 17.2.1 Complex causality (pp. 408–409)#

• Two epistemic cultures (p. 408).
• Field ecologists take a “diversity‑oriented approach”: detailed case studies of many system-specific attributes “with configurational complexity” (McGlade 2003).
• Resource managers take a “variable‑oriented approach”: few attributes, many cases.
• The two are “the ends of an inverse relationship between the number of cases and variables”.
• These are “a fundamentally different way of developing counterfactual thinking”. “Their misapplication has contributed to the catastrophic demise of some the world’s key ecosystems and resources.”
• Definition used throughout. Counterfactual thinking = “the dependence of whether, when and how one event occurs on whether, when and how another event occurs and the possible alteration of events” (pp. 407, 408, 424). In use, it means roughly “the causal model or inferential frame an actor brings to the evidence”. The chapter calls it “correct”, “strong”, “flawed” or “new”. The wording echoes the philosophical account of causation as counterfactual “influence”; no source is cited.
• The two cases framed as a contrast (pp. 408–409).
• Keeling (Box 17.1): “correct counterfactual thinking”.
• Cod (Panel 17.1): “a variable‑oriented approach based on flawed counterfactual thinking and a poorly described polythetic scheme”. Here, “false inferences were used to suppress early warnings leading to an unanticipated, sudden collapse and irreversible demise”, and knowledge was split into “disciplinary, community and geographic silos” (p. 409).
• Diversity-oriented analysis (p. 409).
• Causation is “conjunctural, or combinatorial and heterogeneous”.
• It seeks “specific necessary and sufficient conditions” (Hogg et al. 1989).
• It works best “when causation is complex, where no single condition is either necessary or sufficient or when causes are sufficient only in combination”.
• Detecting necessary conditions for change “can have very powerful policy implications”.
• A concession: “Diversity‑oriented analyses tend not to provide insights into causal generalisations, but rather into the validity of specific processes” (p. 409). The approach the chapter favours trades generality for case-specific validity.
• Variable-oriented analysis (p. 409).
• It rests on “an additive linear model” that isolates each variable’s effect “net of the effects of all others”.
• It groups cases into “polythetic” schemes: members can differ on some attributes if they match on most.
• Claim: such schemes “violate the core principles of the configurational or diversity‑oriented approach used for studying resilience”.
• The open question is “whether the approach adopted can make a difference” to anticipating surprise.
• Lineage (not cited): the case- versus variable-oriented contrast, conjunctural causation and “diversity-oriented” research come from Charles Ragin’s comparative and fuzzy-set methods. McGlade 2003 (“a diversity based fuzzy systems approach”) adapts that tradition.
• Keeling as exemplar (p. 409).
• The chapter acknowledges “some sceptics” who invoke solar output. Against them, “the vast majority of policymakers and scientists now agree that the root cause, i.e. human activity, can be seen in the measurements of carbon dioxide”, largely thanks to Keeling.
• The Keeling Curve (cross-referenced to Figure 14.1) is “one of least disputed pieces of evidence of climate change”.

Box 17.1 The Keeling Curve — “how close observation revealed the secret of climate change” (p. 410)#

The narrative (the box is unsourced):

• Keeling started at Mauna Loa “in January 1958”.
• At 27 he told colleagues all existing CO2 measurements “were wrong or misleading”. “Recognising that no one else was really interested in taking consistent, continuous measurements”, he did it himself: first at Caltech, measuring every three hours and camping in remote places, then at Scripps from 1956.
• He ran a global network of “carefully selected sites… including Antarctica” to find a “global base level”.
• On persistence: while others “moved on to discover new phenomena, Keeling decided that he could not give up… the science itself forced him to behave unusually as a scientist”.
• Mauna Loa is described as “a former US military site”.
• He is said to have filled a flask by hand “each day” for 40 years.
• The curve shows “a jagged edge with an amazing regularity and a steep relentless rise”. Overlaying temperature gave him “foreboding in the 1970s”.
• He “faced cuts to his funding” several times “but always found ways to carry on”.
• The analyser was installed March 1958 and decommissioned January 2006. “Charles Keeling died in 2004”; his son Ralph continues the work.

Accuracy flags (background knowledge; to verify): - Keeling died in June 2005, not 2004. - The box gives January 1958 in one place and March 1958 in another; March is the standard date. - Mauna Loa Observatory was a US Weather Bureau station. I know of no military use. - The observatory ran a continuous analyser, and Keeling was based at Scripps. The picture of him personally filling flasks every day is romanticised. - Chapter 14 of this report names him “David Keeling” (index: Keeling, David, pp. 312, 339). His full name was Charles David Keeling.

This celebratory, unsourced box is the chapter’s main positive exemplar. (Minor: it also says “Scripps Institute” for Scripps Institution of Oceanography.)

17.2.2 A tale of two cods (p. 411)#

• The MSY paradigm. A “simple equilibrium model became paradigmatic during the 1960s”: Maximum Sustainable Yield “under different extraction rates and constant rates of birth, growth and death”. It drew on optimal control and the calculus of variations (Hotelling 1931; Pontryagin et al. 1962), techniques “successfully applied in theoretical physics, aeronautics, chemistry and management”. From there it spread to economics and resource policy (Dasgupta 1982; Clark & Kirkwood 1986) (p. 411). My gloss, not the chapter’s wording: an analytical technique carried its equilibrium and constancy assumptions across into a new domain.
• Early warning ignored. “Concerns were raised by some about the utility of MSY models (Larkin, 1977), but these early warning signals were ignored.” Larkin’s title was “Epitaph for the concept of maximum sustainable yield” (p. 411; ref. p. 427).
• Dating problem. The text says “the northern cod… collapsed in 1986” (p. 411), and says that only then did “questions [begin] to be raised” about MSY, despite Larkin’s 1977 critique cited just above. The chapter is internally loose on the date. The panel speaks of “the collapse of the northern cod… in the 1980s” (p. 412) but has the minister closing the fishery in 1992, “saying it had collapsed” (p. 414). The report’s own Annex 2 (p. 703) says “1992 The stock collapses and a moratorium is imposed”. 1986 is the year of the first specific external warning (the Keats/Memorial review).
• The Canada–Norway contrast.
• Norway’s “immediate reaction… was to impose a moratorium and to invest in more research”. Three years later, Barents Sea cod spawning biomass “was bigger than it had been in twenty‑five years, whilst the potential collapse of Canadian fish stock was still being covered up” (p. 411). “Covered up” is unsourced.
• Norway’s Fisheries Minister “Oddrum [sic] Pettersen” is reported (in indirect speech, not a direct quotation, and with no source) as saying that “although the warning signs were there, the knowledge to understand them was missing”. Hence “precautionary actions through the moratorium” plus research on cod–capelin–mammal links (p. 411). Flag: Oddrun Pettersen appears to have become minister in late 1990, after the January 1990 measures. To verify.
• The Norwegian causal account (p. 411).
• Large year-classes of 2–3-year-old cod ate “over a million tonnes of capelin” while capelin was also overfished.
• The capelin collapsed; cod turned cannibal; seals switched to cod.
• A change in ocean temperature “triggered collapses” across Russian, Icelandic, Swedish, Finnish, Faeroese and Lofoten coastal stocks.
• Environmental drivers are accepted in Norway’s story. The Canadian equivalent is called “myth-like” on the same page (17.2.3).
• Burden of proof (p. 411).
• Canadian scientists insisted there was “no proven connection between fishing on spawning stocks and the survival of the 0‑year class”. Officials said there was “no evidence that fishing on spawning grounds in any way harmed the stock”.
• In Norway “such fishing was banned” because scientists “were convinced that there was such a relation” even though it was hard to show.
• The chapter’s summary: “The Norwegians had understood that no evidence of harm is not the same as evidence of no harm.”
• Flag: Norway’s traditional Lofoten fishery targets spawning cod. The chapter may mean trawling restrictions. To verify, because this is the chapter’s cleanest precautionary example.

Panel 17.1 “The last hunters” — Jacqueline McGlade, first person (pp. 412–414)#

A view from inside (p. 412)

• In 1980 McGlade was a new DFO scientist assessing pollock. She describes DFO as “the largest employer of biologists in the public service”, with more than 2,200 staff and nearly one third of the departmental budget. The offshore zone was a third of Canada’s territory.
• Toolkit: a small team, a computer, historical data, and a “Virtual Population Model” derived from Beverton–Holt. Also winter ship-time for recruitment surveys, juvenile tagging and genetics.
• Target: F0.1, the fishing mortality at which the marginal yield of extra effort is 10% of that on a lightly exploited stock.
• Fishers’ knowledge. Talking to fishermen “was not a common practice, nor was it a regular part of the assessment”.
• CAFSAC and “tuning”. At the annual CAFSAC review, “you had to leave your ego at the door, because you would often be sent out again and again to redo the analysis using new combinations of recruitment, fishing effort and mortality rates: this was the reason why it was called ‘tuning’”. This describes iterative fitting under committee pressure. The panel does not allege tuning to a predetermined answer.
• Known weaknesses. There was a shared worry “that the models were too simple”. Data were “often out of date”, and fishers’ logs were unreliable. Meanwhile “successive ministers would say that science was the foundation of everything”.

Canada’s northern cod (pp. 412–414)

• Sources and question. Named sources: Finlayson 1994, Michael Harris 1998 (a journalist) and Steele et al. 1992. The question: “why, with such open support and well funded research, did DFO scientists not anticipate the problem?” (p. 412).
• 1980s, fishers’ warning. Inshore fishers saw condition deteriorating. “The industry” asked Memorial University biologists to review the 1986 assessment. (The panel does not say which part of the industry; Annex 2, p. 703, calls it the “Keats Report commissioned by inshore fishers”.) They found DFO “had seriously overestimated the size of the stock” and linked offshore landings to poor inshore catches. “The surprise was that they had used the same data and come out with startlingly different conclusions” (p. 412).
• 1986, internal inconsistency. The CAFSAC document projected biomass growth even though TACs “had been at least twice as high as they should have been over the previous eight years”. When surveys showed “an enormous abundance of cod”, scientists “should have realised that there was no biological means” for such an increase with “so few fish inshore” (pp. 412–413).
• The 1960s Templeman warning. The panel offers this as “one potential explanation for the increase” shown by the surveys. Overfished spawning schools “would continue to concentrate but in a smaller areas. Thus catch rates might continue to increase even as the stock was collapsing. Unfortunately, DFO accepted the data as they stood” (p. 413). This is the hyperaggregation mechanism. Note the panel presents it as a potential explanation. It gives no documentation that commercial or survey catch rates were in fact inflated this way.
• 1987, Alverson review. The panel used DFO data plus fishermen’s “anecdotal evidence” and found DFO “had got its figures tragically wrong”. The causes were “errors in interpretation of some of the survey information, compounded by a reliance on a faulty mathematical model”, magnified because “it had taken several years to find and admit to the problem”. Alverson blamed a government “reticent in making effective and timely decisions” and DFO’s thin knowledge of cod biology and ecosystem (p. 413). Annex 2 dates it to 1988.
• DFO’s alternative explanation. DFO then proposed that the missing inshore cod “were trapped in warmer water, avoiding the cold intermediate layer, rather than being fished out” (p. 413).
• 1988, TAC raised. The Alverson report was the first public airing of concern. Even so, Ottawa, “convinced that the ocean would continue to provide fish”, raised the TAC for 1988. Concerns “were put far behind socio-economic considerations” (p. 413).
• December 1988, quota cut rejected. DFO scientists, “in a dramatic reversal of advice”, proposed a 50% cut for 1989. The federal and Newfoundland ministers refused, fearing “thousands being out of work and bankruptcies everywhere”, and justified this “by saying that the scientists had been wrong before”. Some senior DFO scientists resisted admitting mistakes (p. 413).
• Official line (Assistant Deputy Minister; Parsons 1993): the “sudden, unexpected decline… during 1991 was the result not of high fishing mortality, but rather abnormal environmental conditions … which, through mechanisms not yet understood, may have led to an abrupt increase in natural mortality in the early 1980s” (p. 413). “Early 1980s” is as printed (checked in the PDF). It sits oddly with a decline “during 1991” and may be a slip for “early 1990s”. Unverified against Parsons.
• 1989, more evidence. “Despite the increasing evidence that the unthinkable was happening… the politicians asked for more evidence.” The Northern Cod Review Panel (Leslie Harris) began work. “Meanwhile” (no year given) inshore fishers sued “to stop destructive fishing practices on the spawning grounds. They lost”, and DFO officials kept to the line that there was “no evidence that fishing on spawning grounds in any way harmed the stock”. Ottawa was subsidising Newfoundland to “2.56 billion Canadian dollars” (unsourced; period unclear) (p. 413).
• 1990, Harris report.
• The 1990 quota cuts would “contribute to a further decline”.
• DFO scientists were “lulled by false data signals and, to some extent, overconfident of the validity of their predictions”. They failed to recognise “the statistical inadequacies in their bulk biomass model” and the risk of advice built on “short and unreliable data series”.
• “If there had not been such a strong emotional and intellectual commitment to the notion that the F0.1 strategy was working, the open and increasing scepticism of inshore fishermen might have been recognised as a warning flag.”
• Recommendations: external peer review; a warning that “the models had become more important than the study of the species themselves”; “a massive research effort” (p. 413).
• Ottawa accepted nearly all of them, “but reducing the total allowable catch on cod was not approved” (p. 414).
• 1992, closure. The fishery was closed, “saying it had collapsed due to unusual ecological and environmental conditions”. The panel calls it “one of the most catastrophic failures in fisheries”; “many communities remain devastated”. The panel states that “DFO now admits that it had evidence as early as 1986” and, citing the Auditor General (1997), that “the model used to determine the status of most of the key groundfish stocks has consistently overstated their abundance and understated the level of mortality” (p. 414). Attribution care: only the model quotation is clearly the Auditor General’s. The “evidence as early as 1986” admission is the panel’s own statement, with no separate source and no date for “now”. The Auditor General’s finding covers “most of the key groundfish stocks”, not northern cod alone.

Barents Sea cod (p. 414)

• 1989 survey. Autumn 1989: stock “at their lowest for over a hundred years”. Scientists and fishermen were already worried about small fish and inflated estimates. Cod quotas were “slashed”, the capelin fishery closed, and fishers reported fish “appeared to be starving”.
• Political response. Quoting Michael Harris (1998): “When scientists told the Norwegian politicians that there had to be severe quota cuts, the politicians acted immediately and backed the cuts despite a firestorm of criticism from their constituents”. Oslo also cut the number of fishermen and vessels.
• January 1990 “moratorium”. Fishers’ incomes fell by up to 40%. The state “in the end” spent more than EUR 50 million removing capacity. The panel attributes recovery to “the prompt action by the government”: signs appeared by 1991, and by 1992 spawning biomass was “larger than it had been for 25 years”. The causal attribution is the author’s; no stock-recruitment or environmental analysis is offered.
• Flag: whether this was a true cod “moratorium”, as opposed to deep TAC cuts, vessel quotas and early closures, needs checking. The source is journalistic.

17.2.3 Post moratorium (main text, pp. 411, 414–415)#

• The general question. What lessons follow, given “the majority of the world’s commercial fisheries are over‑exploited and many resources are managed using similar models”? “Fisheries management is like a black‑box” (p. 411).
• Official blame. Minister John Crosbie blamed three things: stock overestimation, “foreign overfishing” and “devastating ecological factors”. “The effects of ecological factors became myth‑like” (p. 411).
• Three flaws, following Finlayson & McCay (1998): 1. One stock unit. The stock was treated “as one unit whereas it was known to be comprised of distinct populations with different migratory behaviour”. Not including this variability “may have contributed to the overestimation of biomass” (deYoung & Rose 1993; Finlayson 1994) (p. 411). Note the hedge. 2. Constant parameters. “Key variables, such as recruitment and natural mortality, were treated as constants”, and poorly chosen averaging periods compounded the overestimates (p. 414). 3. Distorted and biased data. Data were “seriously distorted by the practice of discarding undersized fish”. Offshore landings were favoured to “tune” the model. Inshore landings were excluded because of difficulties obtaining consistent “‘catch per unit effort’” and “messy and often anecdotal data”. So “evidence of decline in the abundance and size of inshore migrating fish was apparently missed”, feeding “DFO’s inclination to dismiss the concerns of inshore fishers” (Steele et al. 1992; Neis 1992) (p. 414).
• Incentives. Basic cod biology was neglected partly because of “the career structure of fisheries scientists who were assessed on their publications in peer reviewed journals”. Northern cod papers “fell from more than 10 per year to 2 annually over a twenty year period”, while Norway “actively encouraged” research (p. 414). The figures are unsourced and the causal logic is compressed.
• Hubris and reassurance. “The hubris amongst senior government officials in denying that there was anything wrong… was a projection of the government’s own crisis around its institutional authority and epistemological legitimacy”. Example: the February 1988 DFO Fo’c’sle article “The Science of Cod”. In it the Director General called it “reassuring” that the Alverson Task Group and CAFSAC agreed, said “The credibility of DFO scientific advice was not questioned”, and said the stock “continues to increase” (p. 414).
• Silence and minimisation (pp. 414–415).
• Scientists with misgivings “as in any civil service… were not allowed to speak publicly about the issues once it had become policy”.
• Many dealt with “‘paper fish’”.
• Presentations to advisory bodies tended to “minimise the problems associated with data reliability, uncertainty, the models… or even admitting to the possibility of unknown unknowns!” (footnote quoting Rumsfeld 2002).

Table 17.1 Pedigree of knowledge (p. 415; credited to McGlade 2002)#

This adapts the Funtowicz–Ravetz NUSAP pedigree matrix (my identification). Both were on the report’s advisory board (p. 5) but are not credited. The ∞ row is an addition. From background knowledge (unverified): in the standard matrix, “peer acceptance” at level 2 is “Medium”, so the table’s repeated “Low” at levels 2 and 1 may be a slip. The table supports the claim that causal knowledge in socio-ecological systems has “low pedigree” (p. 415). Nothing, including the cod assessment, is actually scored against it.

17.2.4 Knowledge systems (pp. 415–416)#

• Why the models persist. “Simple variable‑oriented analyses are no match for the complexity of real socio‑ecological systems.” They survive partly because of “the lack of consensus about which approaches could replace them”, rooted in low causal pedigree (p. 415). An important concession.
• Silos. They produce Yaffee’s “recurring nightmares”. “In situations where strong counterfactual thinking is needed, for example in developing early warnings, the presence of silos of knowledge can become a real hindrance” (p. 415).
• Unfamiliar expertise. Interdisciplinary inputs from fields planners don’t know dilute quality control. Organisations face “a chaotic mixture of information, analysis and interpretation with no paradigmatic structure” for combining scientific, interdisciplinary and indigenous knowledge (p. 415).
• “Second-order science.” Individuals “must rely on other peer groups and experts to be able to evaluate the information within their own domains of expertise” (p. 415).
• Broadening expertise. Direct scientific evidence is often missing, while local and lay knowledge is rich. So “the concept of an expert as part of the system of governance has to be broadened to include those who have particular knowledge about a system but who do not necessarily walk the corridors of power and/or of science” (p. 416).
• IPBES recognises indigenous and local knowledge “yet it is still struggling to find operational ways to include such knowledge” (p. 416).

17.3 Planetary dynamics; 17.3.1 Navigating the Anthropocene (pp. 416–417)#

• Four concepts from non-linear dynamics: planetary boundaries (Rockström et al. 2009), tipping points (Lenton et al. 2008; Schellnhuber 2009), panarchy (Gunderson & Holling) and resilience (Walker & Salt 2006). The question is whether they help anticipate sudden change (p. 416).
• The Anthropocene. Systems are changing faster than at any time in the Holocene. Evidence has been “sufficient for geologists to conclude that an epoch‑scale boundary has been crossed” (Steffen et al. 2011; Zalasiewicz et al. 2011). This overstates how settled the matter was. The chapter’s own cited Zalasiewicz et al. (2011) paper is titled “The Anthropocene: a new epoch of geological time?”, with a question mark (ref. p. 428). Threats listed run from sudden-onset (spills, accidents, quakes, storms, epidemics) to slow-onset (air quality, drought, food security) (UNEP 2012) (p. 416).
• Theory ran ahead of practice. Non-linear and chaos theory developed in the 1980s–90s (May 1982; McGlade 1999), but “few field biologists could really make use of the ideas beyond concepts and definitions”. Climate science under the IPCC carried the advances instead (p. 416).
• Planetary boundaries. Nine limits are listed. “None of these limits can be entirely dismissed… or confirmed”. Because of scarce in situ validation and “the inbuilt latency of the planetary system”, “there will be many decades of tantalising ignorance about whether or not any of these boundaries per se have been crossed” (pp. 416–417).
• Tipping elements will “more likely” define future surprises. Some are “already” under way: Arctic sea ice, boreal dieback, Greenland, West Antarctic instability, Atlantic deep water formation, permafrost. Others are potential: a climate-induced ozone hole, a greener Sahara, Indian and West African monsoon shifts, ENSO changes, Amazon dieback, Antarctic bottom water (p. 417).
• The epistemic problem. Emergent properties “(and consequently early warning signals)” arise across scales and are “observed by scientists across many disciplinary silos, which are themselves populated by elites and people with their own paradigms and language” (p. 417).
• Prescription. “Identifying thresholds based on a generic model of tipping points is an exercise in futility”. Instead, use “cascading thresholds”, with early warnings “linked to observations of key processes on a local and regional scale” (p. 417).

17.3.2 Adaptive cycles and panarchy (pp. 417–419); Table 17.2 (p. 418)#

• Resilience defined. The Resilience Alliance’s research programme. Holling (1973) distinguished “engineering resilience” (speed of return) from “ecological resilience” (how far a system can be perturbed “without experiencing a regime shift”) (p. 417).
• Hidden cycles. Adaptive cycles emerge from ecosystem–people interlinkages, and “these, sometimes hidden, cycles” are what managers actually face. “Alternative state” is “a misnomer”. Systems are better described as “configurations of states” with alternate regimes separated by thresholds, as in the Sahel and “permanent summer algal blooms in the Baltic Sea” (p. 417).
• Walker et al. (2006) heuristics. Adaptive cycles and panarchy, plus resilience, adaptability and transformability. The cycle runs r (growth), K (conservation), Ω (release: bound-up resources freed, structure collapses) and α (reorganisation: “novel structures can take hold”). Examples: fire, spruce budworm, algal blooms, drought (pp. 417–419).
• Table 17.2 (adapted from Walker et al. 2006) gives seven thumbnails:
• Causse Méjan (France): grazing-dependent diversity; pine invasion; resilience via farming culture.
• Caribbean reefs (Hughes): loss of the urchin grazer “in the 9170s [sic]” after disease, then persistent algal blooms; no-take areas unlikely “to stem region-wide loss of resilience”. The Diadema die-off is usually dated 1983–84.
• Southern Madagascar spiny forest (Elmqvist): informal institutions protect patches; insecure land rights.
• Everglades (Gunderson): “a perversely resilient social system based on conflict and a formal closed network of government agencies”.
• Gorongosa (Mozambique): wildlife “decimated in the war of Independence in the 1980s”; 10,000 residents without services who “provide local protection”. Background (unverified): Mozambique’s independence war ended in 1974–75; the 1980s conflict was the civil war.
• Goulburn-Broken (Australia): more than 70% of vegetation cleared; salinity; “costs of maintaining the regime mounting”, economy “more brittle”.
• Northern Highland Lakes (Wisconsin): invasives; management “slow to adjust”; a possible “non‑reversible threshold”.
• These are illustrative only and are not analysed.

17.4 Institutional fit and resilience; 17.4.1 Institutional fit and governance (pp. 419–423)#

• Institutional fit. Poor fit confounds stewardship and leads to “unexpected surprises, excessive rent seeking and high transaction costs” (Ostrom 2005; Folke et al. 2007; Kofinas 2009). Its causes include “mere folly (Tuchman, 1984), complexity, uncertainty about roles, a lack of ownership and property rights, institutional frailty and vested interests” (p. 419).
• Trust and framing. “The loss of trust in bureaucracies has been growing.” Debates polarise around “false dichotomies, such as choosing between protecting the environment and employment or energy, or diversionary arguments such as the tragedy of the commons”. For northern cod, “the commons were neither open‑access or unregulated, but the use of this discourse meant that the only solution sought was downsizing of the fleets, rather than a closer examination of… flawed scientific advice, vested interests and government priorities” (p. 419).
• Who changed after the collapse (p. 419).
• The chapter asks what happens in a real crisis: do “those institutions with most to lose become destabilised or do they attempt to renegotiate their power relations? Do those with most to gain seize the opportunity for restructuring?”
• “It has been the social structure of the fishery that has collapsed even though the inshore fishermen potentially had the most to gain.”
• Scientists and managers made “only… modest changes despite being at the core of the problem”.
• “This might suggest that modern science is much stronger and more deeply embedded as a social and political authority than generally assumed”. It “will take more than the collapse of a resource… to displace the existing form of authority that has been built on a certain control of science”.
• An original claim, but asserted: no evidence on post-1992 reforms is given.
• Nunavut turbot case (Nunavut Tunngavik Inc. v. Canada, [1998] 4 F.C. 405) (pp. 419–420):
• The Minister raised the turbot quota “against the advice of the NAFO scientific council” and gave the increase “to his own constituents rather than to the Inuit of the Nunavut region or Denmark”.
• An internal memo warned Canada “could be seen as hypocritical” for exhorting others to share the conservation burden.
• Judge Douglas Campbell set the decision aside.
• On appeal the Minister claimed “‘absolute discretion’”, exercised under “‘numerous fluctuating policy concerns which go beyond… conservation… to include cultural, political, scientific, technical and socio‑economic considerations’”.
• The appellate court (which the chapter calls the “Court of Judicial Review”) rejected the appeal as the Minister “had not acted lawfully”. It insisted, however, that it was “‘not to become an academy of science to arbitrate conflicting scientific predictions’” and would review legality, “‘not its opportunity, wisdom or soundness’”.
• Chapter’s diagnosis: no “legal framework in which the Minister’s actions could be judged on the basis of evidence of fairness and long‑term sustainability” (p. 420).
• I could not retrieve the judgment to check who is speaking in the “absolute discretion” passage, or the ground of unlawfulness. Background knowledge (unverified): “in his absolute discretion… issue… leases and licences for fisheries or fishing” is the wording of s. 7 of Canada’s Fisheries Act, and the “common property resource” rationale echoes the Supreme Court’s Comeau’s Sea Foods (1997) judgment. The “academy of science… legislative upper chamber” line also appears to be quoted from earlier Federal Court authority (Vancouver Island Peace Society, 1992). So the passage the chapter presents as the Minister’s own claim may be the court’s statement of settled law, which would weaken its use as evidence of ministerial attitude.
• Bridging institutions. The CBD, the straddling stocks agreement, the Millennium Ecosystem Assessment, IPBES, TEEB and ABS. Even so, “even with highly cohesive, well‑resourced institutions the consequences of a poor institutional fit remain one of the largest obstacle to the resilience of our fisheries” (p. 420).
• Governance types (McGlade 2001; the typology and the phrase “hollow out the state” match R.A.W. Rhodes, uncited): minimal state, corporate governance, new public management, good governance, sociocybernetic systems and self-organising networks. The networks “can resist government steering”, and can “hollow out the state through privatisation and by limiting the scope of public intervention”. “Swarms of amorphous groups” run single-issue campaigns (dolphin-friendly tuna, anti-sealing and whaling). Governance is “increasingly operative… legitimacy is more ambiguous” (p. 420).
• Consensual versus instrumental management. Communities that manage “through consensual actions and norms, such as religion, deity or faith”, reportedly “learn more quickly and show a greater capacity to adapt and even transform” in crisis (p. 420). Asserted; no evidence given.
• Goulburn-Broken. After the 1970s water-table crisis, networks, leaders and committees pooled knowledge and lobbied, “Despite all this, government leaders opted for adaptation rather than transformation” (p. 420).
• Everglades. A government-dominated network shifted to government plus NGOs, “which opted for transformation”, with new institutions and technology “on the back of significant levels of investment” (pp. 420, 423). This sits awkwardly with Table 17.2’s “perversely resilient social system based on conflict and a formal closed network of government agencies” (p. 418). The table may describe the earlier, government-dominated phase, but the chapter does not reconcile the two.
• Conclusion. “Two of the most critical factors for adapting to change and transformation are institutional fit and leadership” (p. 423).

Panel 17.2 Knowledge-based management of EU fisheries — Constança Belchior and Johnny Reker (pp. 421–422)#

• The paradox. The EU runs “one of the most complex and expensive science-based management systems in the world”, yet since 1983 stocks have “continued to worsen”. 72% of assessed EU stocks are overexploited and “over 20% fished beyond safe biological limits”. NE Atlantic overfished stocks fell “from 94 % to 63 % from 2004 to 2010”, against 82% in the Mediterranean (p. 421). The NE Atlantic trend is an improvement, in tension with “continued to worsen” (a claim the panel sources to O’Leary et al. 2011 and “EU, 2011”).
• Data gaps. Adequate data are missing for “approximately two-thirds” of NE Atlantic TAC stocks. In 2010 only 16 of 102 candidate Mediterranean species could be assessed. Member States have failed their legal data obligations, “introducing greater uncertainty into the scientific advice, and thus further undermining its credibility” (p. 421).
• Figure 17.1 (NE Atlantic stocks, 2005–2011, EEA based on EC 2011a; approximate readings from the chart, re-checked in the PDF). Four stacked categories:
• No advice: about 31% (2006) to 42% (2005, 2011); about 35–42% in most years.
• Advice other than MSY: about 20–44%, peaking in 2006.
• Overfished: about 35% (2005), dipping to about 23% (2006), about 31% (2007–09), about 27% (2010) and about 23% (2011).
• Fished at MSY: rose from about 2% (2005–07) to about 14% (2011).
• My reading: overfished plus fished-at-MSY together stay at roughly 35–38% of stocks throughout. So the MSY gain largely mirrors the fall in “overfished” within the minority of stocks that have MSY-based status, which fits the “two-thirds” data gap.
• The panel says the figure “was intended to show fishing opportunities, but inconsistencies in the data… point to the underlying difficulty of producing coherent assessments” (p. 421).
• Why management has failed. Citing several analyses, the panel says the main conclusion, “alongside the lack of scientific knowledge and ineffective management options”, is “the very nature of the decision-making process” (p. 421). So the panel itself concedes that knowledge gaps and weak management tools contribute, not just political override.
• Advice overruled. “Scientific advice has been largely overruled”. “Between 1987 and 2006, the TACs only matched scientific advice in 8 % of the assessed stocks” (Piet et al. 2010). Since 2003, TACs have exceeded advice “by an average of 47 %” (EC 2011a) (p. 421).
• Impacts hedged. “The actual impacts of such mismatches remains uncertain” (O’Leary et al. 2011), but are “likely to be significant” given cumulative overfishing and increased capacity (p. 421).
• Seabed and biodiversity. Benthic recovery in low-natural-disturbance areas of the Greater North Sea takes “7.5–15 years following only one pass of a beam trawl” (OSPAR 2010). Some German areas have had up to 4,000 hours of beam trawling. In the context of “European marine biodiversity”, only 2% of the species and 10% of the habitats reported under the Habitats Directive are in favourable conservation status (HELCOM 2010; EEA 2010) (p. 421).
• Ecosystem approach. An ecosystem-based approach “has become one of the CFP’s objectives under the EU Marine Strategy Framework Directive”, meant to account for hidden long-term costs, externalities and subsidies (p. 422).
• The economic case (NEF: Crilly 2012; the reference list calls it “New Economics Forum”). Restoring 43 stocks to MSY would add 3.53 million tonnes and “EUR 3 188 billion annually” (evidently EUR 3.188 billion). That is more than five times the annual subsidies and could support more than 80,000 jobs. The EU imports more than 65% of its seafood. The World Bank’s “sunken billions” estimate is USD 50 billion a year globally (p. 422).
• Knowledge is enough. “Thus it seems that there is sufficient available knowledge to underpin informed decision-making to support an ecosystem approach” (p. 422). This sits in some tension with the panel’s own emphasis on data gaps.
• Diagnosis. Nothing will change “as long as decision-making in the CFP allows for short-term economic interests to compete with and outcompete long-term sustainable and precautionary scientific advice at the very beginning”. Asked whether Europe has learned, the panel answers: “The short answer is yes, but we are still not reacting to our knowledge fast enough” (p. 422).
• The reform. The 2011 CFP proposal moves from “‘based on scientific advice’” to “‘in accordance with the best available scientific advice’”, gives stakeholders more responsibility, and integrates the precautionary and ecosystem approaches. Success is “questionable” while science is undermined by non-compliance or “narrow political interests and negotiations”, a point the panel says the European Parliament also recognised (EP 2012). The key challenge is “the level of respect for scientific advice as a fixed boundary condition for natural capital when negotiating political agreements”. Science-grounded management is “undoubtedly a more objective way”, “unlike the vested political decision-making system that prevails in the CFP today”; the panel calls for transparency, inclusiveness and “broader representative constituencies” (p. 422).

17.5 Future prospects (pp. 423–424); Table 17.3; Box 17.2#

• The six lessons (p. 423): 1. Close observation. 2. A diversity-oriented approach “that reflects the complex causality of the real‑world context”. 3. Widening knowledge sources. 4. “Reducing delays between early warnings and actions”. 5. Understanding “the full adaptive cycle”. 6. Building “learning, leadership and innovative capacities… to enable transformation rather than only adaptation”, with “greater participation and transparency regarding vested interests”.
• Explicit advice and triggers. Advice must be “far more explicit about the dynamics of processes across different time and space scales and uncertainty”. Managers and communities must build “consensus — and where not possible compromises (van den Hove, 2006) — about how to handle uncertainty and which diagnostic criteria and metrics will be used to elicit action in the case of early warning signals”. Examples: water-use thresholds, forest-clearance patch sizes, wildfire management, agistement (p. 423).
• Table 17.3 (p. 423) sets out warning types, examples and responses:

• Institutions under pressure. “The fundamental assumption that the existing institutional landscape is fit for purpose should also be examined carefully.” The chapter cites Hobbes and Machiavelli on participation and power, and a range of regimes from laissez-faire to international (McCay 1993). “Property rights and the authority to constrain exploitation are at the heart of the matter and inevitably come under intense pressure during a crisis” (p. 424).
• The hubris of predictability. Interventions assume “all future states are contained within the present structure”. This “parallels that of the financial world in which a constrained set of tools and those using them (known as ‘quants’)” created “a false protection against unknowable uncertainties”. Finance and resources became “intertwined via market volatility and the mispricing of commodities” in the economic crisis (McGlade 2009) (p. 424).
• Close. “Resilience in nature contains the seeds not only of surprise but also of transformation” (p. 424).
• Box 17.2 Definitions (pp. 424–425). Key entries:
• Adaptive management: approaches that “recognise uncertainty and encourage innovation while fostering resilience”.
• Institutional fit: a match of spatial and temporal “sizes”, as a shoe fits a foot.
• Panarchy: nested adaptive cycles; knowing them helps “identify leverage points”.
• Regime shift; resilience.
• Tipping point: the popular sense after Gladwell, and the formal “tipping element” of Lenton et al. The chapter notes the term’s plural meanings: positive feedback, hysteresis, noise-dependent bifurcation.

References (pp. 425–428): observations#

• Mixed sourcing for the Canadian account. It is anchored in primary and official sources (Alverson 1987; Leslie Harris 1990; Parsons 1993; Auditor General 1997) and in scholarly secondary work (Finlayson 1994; Finlayson & McCay 1998; Steele et al. 1992; Neis 1992). The Norwegian account rests mainly on Michael Harris 1998, a journalist’s “true crime story”. The two different Harrises must not be conflated.
• Uncited references may suggest the chapter was cut from a longer draft: Ferguson 2012, Rogers 1986 (the Challenger commission) and Stokes et al. 1993. (EU Court of Auditors 2011 is probably the “EU, 2011” cited twice in Panel 17.2, p. 421, so it is likely not uncited.)
• Cited but missing from the reference list: May 1982 (p. 416), Holling 1973 and Holling 1996 (pp. 416–417).
• Minor citation slips: Lenton 2008/2009; Gunderson spelled “Gundersen” and dated 2001/2002; Schellnhuber 1999 in Box 17.2 but 1994 in the references; “Villasante et al., 2010” cited for a single-author paper; “Petersen”, “Lynan”; IPBES called “International” (p. 420) and “Intergovernmental” (p. 416).

Case timeline#

A. Northern cod, Newfoundland (Canada)#

Lags. From the first specific, independent scientific warning (1986) to closure (1992): about 6 years. From DFO’s own advice to halve the quota (December 1988) to closure: about 3.5 years. In that period the proposed 50% cut was rejected, the 1990 quota reductions were judged insufficient by Harris, and his call to reduce the TAC was “not approved”. (The TAC increase was for 1988, before the December 1988 advice.) From the paradigm-level critique (Larkin 1977) to closure: about 15 years. From Templeman’s mechanism warning: about 25–30 years.

Harms. “The social and economic consequences have been enormous and many communities remain devastated” (p. 414). The chapter reports the collapse of “the social structure of the fishery” (p. 419). It mentions CAD 2.56 billion in federal subsidies to Newfoundland by about 1989, before the closure (unsourced; period unclear) (p. 413). “Thousands being out of work and bankruptcies everywhere” is what ministers feared from the proposed 1989 quota cut (p. 413), not a reported outcome of the collapse. The chapter gives no systematic figures for job losses or compensation costs.

Who bore the costs. The chapter says “the social structure of the fishery… has collapsed even though the inshore fishermen potentially had the most to gain in the aftermath”, while scientists and managers made “only… modest changes” (p. 419). The inshore fishers’ knowledge was discounted and their lawsuit failed, and offshore landings “tuned” the models (pp. 413–414). The chapter does not break the post-collapse costs down between inshore and offshore sectors.

B. Barents Sea cod (Norway)#

Lag between warning and action: months, measured from the autumn 1989 survey. Scientists and fishermen were “already” worried before that (p. 414).

C. EU fisheries (Panel 17.2)#

• 1983 onward: stocks worsening despite science-based management (p. 421).
• 1987–2006: TACs matched advice for 8% of stocks (p. 421).
• 2003 onward: TACs averaged 47% above advice (p. 421).
• 2002 and 2012 CFP reforms, with the 2011 Commission proposal (pp. 421–422).
• 2004–2010: NE Atlantic overfished share falls from 94% to 63% (p. 421).
• As of 2013: the prediction that reform will fail unless scientific advice becomes a “fixed boundary condition” (p. 422).

D. Nunavut turbot (pp. 419–420)#

• 1997: the Minister raises the turbot quota against NAFO scientific advice and gives the increase to his own constituents.
• Trial judge sets the decision aside. Appeal rejected (reported at [1998] 4 F.C. 405), with the court declaring it will not arbitrate science.

The authors’ own lessons and conclusions#

Lessons the authors derive from their evidence#

1. Method matters. Variable-oriented, equilibrium models (MSY, F0.1, VPA with constant parameters and a single stock unit) were “no match” for socio-ecological complexity. Their continued use “contributed to the catastrophic decline of some resources” (pp. 408, 411, 415). Evidence: cod assessment flaws, pp. 411–415.
2. Early warnings existed and were suppressed or ignored through “false inferences” (p. 409). Evidence: Templeman, Larkin, the inshore fishers, Keats et al., Alverson, the 1988 DFO advice (pp. 411–413).
3. Absence of evidence of harm is not evidence of no harm. Norway acted on a conviction about the stock–recruitment relation without proof, while Canada demanded proof (p. 411).
4. Prompt action with investment in research outperforms delay (Norway versus Canada, pp. 411, 414).
5. Excluding fishers’ knowledge, and the “paper fish” distance of assessment scientists, blinded the institution (pp. 412, 414–415). Hence expertise in governance must be “broadened” (p. 416).
6. Institutional self-defence (hubris, reassurance, civil-service silence, minimising uncertainty) delayed recognition (pp. 413–415).
7. Framing, such as the “tragedy of the commons” and environment-versus-jobs dichotomies, narrowed the solutions considered (p. 419).
8. Science-based authority is resilient to its own failures. The costs of collapse fell on the fishing community, not on the knowledge institutions (p. 419).
9. Legal frameworks can lack any standard for judging long-term sustainability, leaving ministerial discretion unchecked on substance (Nunavut, p. 420).
10. Even rich, science-based institutions fail when political bargaining overrides advice and data obligations are not met (EU; Panel 17.2, pp. 421–422, by the panel authors).
11. Institutional fit and leadership are the critical factors for adaptation and transformation (p. 423). Evidence: Goulburn-Broken and Everglades. Thin.

Recommendations and advocacy#

• The six lessons (p. 423): close observation; a diversity-oriented approach; wider knowledge sources; reduced warning–action delays; understanding the full adaptive cycle; capacity for transformation, participation and transparency on vested interests.
• Management advice should be explicit about cross-scale dynamics and uncertainty. Pre-agree, by consensus or compromise, “which diagnostic criteria and metrics will be used to elicit action” (p. 423).
• Early-warning typology and responses (Table 17.3, p. 423): broad jurisdictional authority, in situ monitoring, multi-scale models, integrated assessments and scenarios, multi-jurisdictional approaches.
• A cascading-thresholds approach rooted in local and regional observation, not generic tipping-point models (p. 417).
• Re-examine the assumption that existing institutions are fit for purpose. Make rights and responsibilities explicit for periods of transformation (p. 424).
• Abandon the hubris of predictability (p. 424).
• Panel 17.2 (Belchior & Reker): treat scientific advice as a “fixed boundary condition”; enforce Member State data obligations; adopt the ecosystem approach and “best available scientific advice”; make decisions more transparent and inclusive; restore stocks to MSY for economic gain (p. 422).
• Harris report recommendations, reported by McGlade: external peer review of assessments and a massive research effort (p. 413).

Mechanisms and dynamics#

1. How the warnings arose#

• From within science, early and at the paradigm level. Templeman’s mechanism warning (1960s) and Larkin’s critique of MSY (1977) came from inside fisheries science itself (pp. 411, 413). The warnings were theoretically grounded but not operationalised in the assessment routine.
• From resource users. Inshore fishermen noticed the deteriorating condition of inshore cod (p. 412), and later analysis confirmed a decline in the “abundance and size of inshore migrating fish” (p. 414). In Norway fishermen reported starving fish (p. 414). The people closest to the resource saw the change first.
• From independent re-analysis. Memorial University’s 1986 review reached “startlingly different conclusions” from the same data (p. 412). Alverson’s 1987 review combined DFO data with fishers’ “anecdotal evidence” and also found the figures “tragically wrong” (p. 413). Both warnings came from outside the assessment institution. My inference: at least in the Memorial case, the difference came from analysis (assumptions or methods) rather than new data. The chapter does not say what the reviewers did differently.
• From internal inconsistency. The 1986 CAFSAC document contained signals that “something was very wrong” (p. 412).

2. How the warnings were contested, filtered or ignored#

• Data-format gatekeeping. Inshore information did not fit the CPUE metric and was “messy and often anecdotal”, so it was not used. Offshore landings tuned the model (p. 414). Templeman’s warning (p. 413) implies that fleets fishing dense offshore spawning aggregations would show inflated catch rates. The chapter does not state this link explicitly. The institution’s standard of admissible data excluded exactly the signal that mattered.
• Indicator hyperstability (offered as a potential explanation). Templeman had warned that catch rates “might continue to increase even as the stock was collapsing”, and the panel offers this as “one potential explanation” for the implausible increase (p. 413). The chapter does not document the catch-rate series itself. Research survey data showing “enormous abundance” were accepted “as they stood” (pp. 412–413). Discarding of undersized fish “seriously distorted” both industry and survey data (p. 414).
• Structural assumptions. A single stock unit and constant recruitment and natural mortality (pp. 411, 414) turned a heterogeneous, changing system into a stable, average one. The model’s assumptions defined what could be seen.
• Emotional and intellectual commitment to the management strategy (Harris 1990, p. 413), and “models… more important than the study of the species” (p. 413). The analytical tool became the object of loyalty.
• Alternative explanations as protection. Cod “trapped in warmer water” (p. 413), “abnormal environmental conditions” (Parsons 1993, p. 413) and Crosbie’s “devastating ecological factors” and “foreign overfishing” (p. 411). The chapter calls ecological explanations “myth-like” (p. 411). My reading: the effect of these explanations was to move responsibility away from domestic fishing. Note, though, that Crosbie’s list did include “overestimation of the stock leading to the setting of total allowable catches that were too high” (p. 411), so the official account was not wholly exculpatory of the assessment.
• Institutional reassurance. The Fo’c’sle article affirmed credibility and continued growth after Alverson (p. 414). The chapter frames this as “a projection of the government’s own crisis around its institutional authority and epistemological legitimacy” (p. 414).
• Muted internal dissent. Civil-service norms kept scientists from speaking publicly “once it had become policy” (p. 415). Uncertainty, stock structure and “unknown unknowns” were played down in advisory presentations (p. 415). Senior scientists were reluctant to admit mistakes (p. 413).
• A credibility trap. Ministers dismissed the December 1988 call for a 50% cut, “supporting their decision by saying that the scientists had been wrong before” (p. 413). My inference: the chapter does not say which past error ministers meant, but the documented error was an optimistic overestimate. So an error in the direction of optimism was used to discount a later, correct warning.
• Demands for more evidence as delay. “The politicians asked for more evidence” (p. 413), and a new review panel was set up while fishing continued.

3. Burden and standard of proof#

• The Canadian position required proof of harm from spawning-ground fishing (“no proven connection”; “no evidence… in any way harmed the stock”). Norway acted on scientific conviction without proof (p. 411). The chapter’s formulation: “no evidence of harm is not the same as evidence of no harm” (p. 411).
• The courts: the inshore fishers lost their spawning-ground case (p. 413). In Nunavut the court would not “arbitrate conflicting scientific predictions” and reviewed only legality (p. 420). Law offered no route for weighing evidence about sustainability.

4. Economics, interests and distribution#

• Socio-economic considerations were placed “far” ahead of stock health. The TAC was raised in 1988, and a cut was rejected in 1989 because of feared unemployment and bankruptcies (p. 413).
• Ministerial discretion openly includes “cultural, political… socio‑economic considerations” (p. 420; see the attribution caveat under 17.4.1). In Nunavut, the additional quota went to the Minister’s “own constituents rather than to the Inuit of the Nunavut region or Denmark” (p. 419): a distributive dynamic. The chapter does not use the word “patronage”.
• Offshore versus inshore: offshore landings were favoured to “tune” the model, and inshore data were set aside (p. 414). The inshore sector then bore the social collapse (p. 419). The chapter documents the data bias and the distribution of harm. It does not claim the offshore bias was driven by offshore interests; that link is my inference.
• EU: short-term interests “outcompete” advice “at the very beginning” of decision-making (p. 422). TACs averaged 47% above advice (p. 421). Member States withhold data (p. 421). The panel frames the economics as a loss, with foregone landings, jobs and “sunken billions” (p. 422). On this framing, precaution is also economically rational over the longer term.
• Costs of precaution were real and borne by fishers. Norway saw incomes fall by up to 40% and spent more than EUR 50 million on capacity removal (p. 414). The chapter presents these costs as the price of a recovery it attributes to “prompt action”, but it does not weigh costs against benefits or test the attribution.
• After collapse, the costs landed on the inshore community, whose social structure collapsed. The knowledge institutions survived with “modest changes” (p. 419).

5. Research funding and incentives#

• DFO was large and well-funded (p. 412), yet basic cod biology research declined. Papers on northern cod fell from more than 10 a year to 2 (unsourced). The chapter attributes this to publication-based career structures, while Norway “actively encouraged” research (p. 414). Money was not the bottleneck; incentives and routine were.
• Norway responded to crisis by funding research into species interactions (p. 411). That research produced a new causal model: the capelin–cod–seal cascade.

6. Mental models of the actors#

• Assessment scientists thought in terms of equilibrium, constants and additive causation (pp. 409, 411). They were confident in F0.1 (p. 413), dealt with “paper fish” (p. 415), and treated inshore reports as anecdote (p. 414).
• Politicians and senior officials believed “that the ocean would continue to provide fish” (p. 413). They saw socio-economic disruption as the binding constraint (p. 413), and regarded science as “the foundation of everything” rhetorically while overriding it in practice (pp. 412–413).
• The Canadian Minister (Nunavut) held a view of discretion as “absolute”, with conservation as one concern among many (p. 420).
• The Norwegian minister, in the chapter’s reported speech, acknowledged ignorance: “the knowledge to understand them was missing”. The response was precaution plus research (p. 411). The chapter singles out this admission as “interesting from the perspective of Late lessons”. My reading: the contrast in epistemic humility is one of the chapter’s main explanatory contrasts, alongside its method-based (diversity- versus variable-oriented) and institutional-fit explanations.
• Resource management in general assumes predictability: “all future states are contained within the present structure”. The chapter likens this to financial “quants” (p. 424).

7. Framing and language#

• The “tragedy of the commons” framing led to fleet downsizing as the only solution (p. 419). False dichotomies set environment against employment or energy (p. 419).
• “Tuning” (p. 412), “paper fish” (p. 415), and “reassuring” and “credibility… not questioned” in the DFO Director General’s text (p. 414).
• “Sudden, unexpected decline” and “abnormal environmental conditions” (p. 413) cast a gradual, warned-of collapse as an unforeseeable natural event.
• The chapter’s own counter-framing: “last hunters”, “covered up”, “hubris”, “myth-like”, “recurring nightmares”.

8. Complexity, thresholds and irreversibility#

• Collapse came suddenly after a long warning period (“unanticipated, sudden collapse and irreversible demise”, p. 409). In Norway, a trophic cascade (capelin, then cod cannibalism, then seals) combined with a temperature shift (p. 411).
• Regime shifts and alternate regimes (Sahel, Baltic blooms, Caribbean reefs; pp. 417–418). Loss of resilience can make a system “more brittle” and more vulnerable to shocks (Goulburn-Broken, p. 418).
• Planetary-scale latency and sparse observation make boundary transgression unknowable for decades (pp. 416–417). Warning signals are scattered across scales and disciplinary “silos… populated by elites” (p. 417).
• Irreversibility: the chapter asserts the collapse was irreversible (p. 409). Benthic recovery takes 7.5–15 years after a single trawl pass (p. 421). A “non-reversible threshold” is noted in the Wisconsin lakes (p. 418).

9. Lock-in and path dependence#

• Paradigm lock-in. The MSY/F0.1 framework persisted despite warnings from 1977 (p. 411). Even after collapse, “why have these approaches not been replaced?” The answer given is the lack of a consensual replacement and the low pedigree of causal knowledge (p. 415).
• Institutional lock-in. Authority “built on a certain control of science” persisted after the collapse (p. 419).
• Infrastructure lock-in (my framing; Table 17.2 does not use the term). The Everglades are maintained through canals, pumps and levees, with a “need for large-scale, expensive, technologically based solutions”. In the Goulburn-Broken, irrigation infrastructure and clearing have altered the hydrology, and the “costs of maintaining the regime [are] mounting” (p. 418).
• Governance lock-in. Leaders chose “adaptation rather than transformation” (p. 420).

10. Institutional behaviour and governance#

• Poor institutional fit, meaning a scale mismatch between institutions and ecosystems, produces surprise, rent seeking and transaction costs (pp. 407, 419).
• Fragmentation of authority, values and information in silos (p. 415).
• Interdisciplinary inputs weaken quality control, and “second-order science” means reliance on others’ expertise (p. 415).
• Networked, self-organising governance can resist steering and “hollow out the state”. Single-issue “swarms” change how legitimacy is formed (p. 420).
• Consensual and normative community management may learn faster in crisis (asserted, p. 420).
• In a crisis, property rights and the authority to constrain exploitation come under intense pressure (p. 424).

11. Substitutes and alternatives#

• The chapter barely discusses alternatives to fishing. It mentions capacity removal (Norway, p. 414), fleet downsizing (Canada, p. 419), no-take areas (reefs, p. 418) and ecosystem-based management (EU, p. 422). The main “alternative” on offer is epistemic: a different way of knowing.

Transferable insights (technology-neutral)#

1. Indicators generated by the activity under scrutiny can stay reassuring while the underlying system collapses. Templeman warned that catch rates “might continue to increase even as the stock was collapsing”, and the panel offers this as “one potential explanation” for the implausible survey increase (p. 413). Offshore landings were favoured to “tune” the model, and discards distorted industry and survey data (p. 414). Strength: moderate–strong. The mechanism was warned of in advance. The Harris review’s “false data signals” (p. 413) and the Auditor General’s finding that the model “consistently overstated” abundance (p. 414) are consistent with it but do not specifically confirm it. The chapter does not document the catch-rate series. Hyperstability of catch rates is, from background knowledge, widely recognised in fisheries science.
2. Commitment to a model or management strategy can turn it from a tool into an identity, making anomalies and outside scepticism invisible. Harris’s “strong emotional and intellectual commitment” to F0.1; “models had become more important than the study of the species”; “paper fish” (pp. 413, 415). Strength: strong for this case. It is the finding of an independent official review, corroborated by insider testimony. Generalisation is moderate.
3. Admissibility rules for evidence can exclude the earliest and most relevant warnings. Knowledge that doesn’t fit the institution’s metrics, such as “messy and often anecdotal” inshore data, is discarded (p. 414). Talking to fishers was not routine (p. 412). Strength: strong in-case, moderate as a general claim. Neis 1992, Steele et al. 1992 and Harris 1990 all point this way.
4. Independent re-analysis by people outside the assessing institution can surface warnings that the institution’s own analysis misses. Memorial University reached “startlingly different conclusions” from the same data (p. 412); Alverson, using DFO data plus fishers’ evidence, found the figures “tragically wrong” (p. 413); Harris recommended that assessments “be peer reviewed by those not involved in DFO” (p. 413). Strength: moderate–strong. Two independent reviews preceded official recognition by years. The chapter gives no cost information and does not say what the reviewers did differently.
5. Where the burden of proof sits largely determines when action comes. Demanding proof of harm lets harm continue; “no evidence of harm is not the same as evidence of no harm.” Pp. 411, 413. Strength: strong as logic, moderate as empirical contrast. Canada versus Norway is a single, confounded pair, and the Norwegian spawning-ground claim needs verification.
6. Past expert error can be turned against later, correct warnings (a credibility trap). Overconfident optimism erodes the authority that later cautionary advice needs, and decision-makers can pick which errors to remember. “The scientists had been wrong before” (p. 413). Strength: moderate. One well-documented instance; the dynamic is plausible and general.
7. Calls for “more evidence” and new review processes can function as delay while the damaging activity continues. In 1989 politicians asked for more evidence; the Harris panel began while fishing continued; the 1990 cuts were judged insufficient (p. 413). Strength: moderate–strong. It is documented in sequence here and echoes other Late Lessons cases.
8. Institutions facing threats to their legitimacy tend to answer with reassurance and denial rather than disclosure. Fo’c’sle “credibility… not questioned”; “hubris”; resistance to admitting mistakes (pp. 413–414). Strength: moderate. The documents are quoted; the “projection of… crisis around its institutional authority and epistemological legitimacy” is the authors’ interpretation.
9. Hierarchical norms can suppress internal dissent once advice becomes policy, and uncertainty tends to be understated to advisory bodies. P. 415 (“a tendency” to “minimise the problems”). Strength: suggestive–moderate. The evidence is mainly McGlade’s insider testimony (“It is clear from the inside”). Finlayson & McCay are cited as asking the question, not as the source of the answer. There is no direct documentary evidence of suppression.
10. When short-term socio-economic costs are concentrated and visible, political systems repeatedly override long-term risk advice. Canada 1988–90 (p. 413); EU TACs matching advice for only 8% of stocks, 47% above advice on average (p. 421); Nunavut (pp. 419–420). Strength: strong. There are multiple independent data points, including quantified EU evidence.
11. Prompt, politically backed action on strong warning signals, combined with research investment and capacity reduction, can allow recovery. The costs of acting are real and borne by specific groups. Norway 1989–92 (pp. 411, 414). Strength: suggestive–moderate. A single case from journalistic sources, confounded by ecology and recruitment, with the details (“moratorium”, spawning-ground ban) needing checks. Norway also acted on a survey showing a 100-year low, so this is better read as a lesson about prompt response than about precaution under uncertainty.
12. Research incentives and routine can hollow out basic knowledge of the very system being managed, even in large, well-funded agencies. P. 414. Strength: suggestive. The statistics are unsourced and the causal logic compressed; the point is plausible.
13. Framing a problem narrows the set of solutions considered. Casting a regulated resource as an open-access “tragedy of the commons” led to fleet downsizing only; environment-versus-jobs dichotomies (p. 419). Strength: moderate. The argument is well made for one case; the chapter supplies no causal evidence about policy deliberation.
14. After a failure, costs can fall on the resource-dependent community, while knowledge institutions whose authority rests on “a certain control of science” persist with modest change. P. 419 (“the social structure of the fishery… has collapsed”; scientists and managers made “only… modest changes”). Strength: suggestive/asserted. It is insightful, but the chapter gives no evidence on post-collapse institutional reforms.
15. Legal frameworks that grant broad discretion, and courts that decline to weigh evidence, leave long-term sustainability without legal purchase. Nunavut (pp. 419–420). Strength: moderate. One clear, documented case; the general point about judicial deference is well known.
16. Mismatch of scale between governing institutions and the system governed breeds surprise, rent seeking and transaction costs. This is institutional fit (pp. 407, 419, 424). Strength: asserted in this chapter. It rests on the Ostrom and Folke literature rather than on the chapter’s own evidence.
17. Early-warning signals in complex systems are scattered across scales and disciplines. Generic threshold models are unreliable; warnings need anchoring in local and regional observation of key processes, with pre-agreed triggers for action. Pp. 417, 423. Strength: suggestive/asserted. It is reasoned but not demonstrated. The “pre-agreed diagnostic criteria” recommendation (p. 423) is logically strong because it removes the argument about evidence from the moment of crisis.
18. Sustained, meticulous, long-term monitoring, maintained against funding cuts and professional incentives to move on, can produce decisive evidence. Keeling (pp. 409–410). Strength: moderate. The example is robust in general, but Box 17.1 is unsourced and contains factual errors.
19. Poor data compliance by those with reporting duties raises uncertainty, which weakens advice and makes it easier to override: a self-reinforcing loop. EU (p. 421). Strength: moderate. Documented (EC 2011a; Piet et al. 2010).
20. Assuming that all future states are contained in the present structure of a system is a recurring form of expert hubris across domains. P. 424; resource management and finance “quants”. Strength: asserted. An analogy drawn by the authors, not evidenced.
21. Crises can open windows for transformation, but leaders often choose incremental adaptation. Leadership and participatory networks decide which path is taken. Goulburn-Broken versus Everglades (pp. 420, 423). Strength: suggestive. Two thumbnail cases, inconsistently described.
22. In interdisciplinary decision contexts, decision-makers depend on expertise they cannot judge (“second-order science”), which dilutes quality control. P. 415. Strength: asserted. Plausible; no evidence given.

Limitations, contestation and bias check#

Who is speaking. The chapter is co-written by the EEA’s Executive Director and a member of the report’s editorial team. Its main case, Panel 17.1, is a first-person account by the lead author, a former DFO scientist. This gives insider insight: the “tuning”, the “paper fish”, the silence of civil-service scientists. It also means:

• There is no independent authorial check on the Canadian narrative.
• The author’s professional positioning (her own diversity-oriented, fuzzy-systems methods) is built into the chapter’s diagnosis.
• No DFO, industry or Canadian government perspective is given in its own voice. Their statements appear only as evidence of error.

The “last hunters” title and phrases such as “covered up”, “hubris” and “myth-like” signal advocacy.

Source quality. The Canadian account draws on strong primary and scholarly sources: Alverson, Leslie Harris, the Auditor General, Finlayson, Steele et al. and Neis. The Norwegian comparison, however, relies mainly on one journalist’s book (Michael Harris 1998) plus an unsourced ministerial remark given in reported speech. Box 17.1 has no sources at all. Several statistics are unsourced: the publication decline, the CAD 2.56 billion, the EUR 50 million.

Factual errors and inconsistencies (in-report or from background knowledge; to verify):

• The main text says northern cod “collapsed in 1986” (p. 411); the panel refers loosely to “the collapse… in the 1980s” (p. 412) but has the fishery closed in 1992 (p. 414); Annex 2 dates the collapse and moratorium to 1992 (p. 703).
• Alverson report: 1987 (p. 413) versus 1988 (Annex, p. 703).
• The Parsons (1993) “official line” quote refers to a decline “during 1991” caused by a rise in natural mortality “in the early 1980s” (p. 413). This is as printed, and may be a slip for 1990s.
• The panel’s “DFO now admits that it had evidence as early as 1986” is easily read as an Auditor General finding. Only the quote about the model (covering “most of the key groundfish stocks”) is clearly the Auditor General’s (p. 414).
• Gorongosa wildlife “decimated in the war of Independence in the 1980s” (Table 17.2, p. 418). From background knowledge, the 1980s war was Mozambique’s civil war.
• The panel describes the VPA as “derived from the equilibrium model of Beverton and Holt” (p. 412). From background knowledge, virtual population analysis is a separate cohort-reconstruction method often used alongside Beverton–Holt yield-per-recruit analysis (unverified; minor).
• Keeling’s death given as 2004 (actually 2005); start date January versus March 1958; “former US military site”; the picture of Keeling personally sampling flasks daily (p. 410).
• The Norwegian Fisheries Minister named (“Oddrum Pettersen”) appears to have taken office after the January 1990 measures (p. 411).
• Norway’s “moratorium” and its “ban” on spawning-ground fishing (pp. 411, 414) may overstate what was done, given the long-standing Lofoten fishery on spawning cod.
• “9170s” (p. 418); “EUR 3 188 billion” (p. 422); “Court of Judicial Review” (p. 420); IPBES named inconsistently (pp. 416, 420).
• The Anthropocene described as settled by geologists (p. 416) when formal ratification was pending (and was later rejected, per my background knowledge). The chapter’s own source, Zalasiewicz et al. 2011, poses it as a question.

None of these errors overturns the core Canadian narrative. They matter, though, because the Norwegian and Keeling “success” exemplars carry much of the argumentative weight and are the least carefully sourced.

Conceptual looseness.

• “Counterfactual thinking” is used idiosyncratically, to mean roughly “causal model”. Calling Keeling’s or Norway’s reasoning “correct counterfactual thinking” is largely a hindsight label for reasoning that turned out right.
• The diversity/variable dichotomy is asserted more than shown. Most of the cod failures identified could, in principle, be corrected within a better variable-oriented, single-species assessment: treating the stock as one unit, holding parameters constant, discard bias, CPUE bias, overconfidence, and political override. Harris recommended exactly that route (external peer review, more research). The claim that the type of approach caused the collapse is not demonstrated. The chapter itself concedes there was no consensus alternative (p. 415).
• Keeling is a poor fit for “diversity-oriented, configurational” analysis. He measured one variable with extreme precision, choosing remote sites to remove local context and obtain a global background signal. That is arguably the opposite of configurational case analysis. The real lesson of Keeling is about sustained measurement standards and persistence.
• The framing of Keeling in the main text is loose. It suggests CO2 measurements show that “the root cause… can be seen in the measurements of carbon dioxide” (p. 409). The curve shows rising CO2; attributing warming requires more than that curve.
• The four-concept lens is not applied to the cases. The abstract promises that planetary boundaries, tipping points, panarchy and resilience provide “an analytical lens through which to explore why many of the warning signals were not seen” (p. 407). In practice section 17.3 is a literature overview, and the cod case is never analysed in adaptive-cycle or threshold terms. The forest, savannah and aquatic “case studies” appear only as table rows.

Asymmetric treatment of environmental causes. Environmental drivers (a temperature shift, a trophic cascade) are accepted as part of Norway’s story (p. 411). In Canada, ecological explanations are dismissed as “myth-like” (p. 411). Background knowledge, to check: the dominant scholarly view is that overfishing was the primary cause of the northern cod collapse. Cold ocean conditions and poor fish condition probably contributed to the timing and to the slow recovery. The chapter is broadly aligned with the majority view, but its treatment leaves no room for genuine scientific uncertainty about environmental factors. That matters for the chapter’s own lesson about humility under complexity.

Case selection and comparison.

• One pair, many confounders. Canada versus Norway is a single pair, with differences in ecology, recruitment luck, stock productivity, bilateral management with Russia and political culture. The chapter does not consider Norway’s own failures, such as the 1960s collapse of Atlanto-Scandian herring (background knowledge), even though it notes capelin overfishing (p. 411).
• The “success” was reactive. Norway acted after a survey showed a 100-year low. That is prompt response to strong evidence, not precaution under weak evidence.
• Hindsight. Claims that scientists “should have realised” (p. 412) sit uneasily with the Norwegian minister’s reported admission that “the knowledge to understand them was missing” (p. 411). The chapter does not ask how, ex ante, one would separate the inshore signal from noise when the offshore survey signalled “enormous abundance” (p. 412).

Internal tension about science’s authority. Panel 17.2 (Belchior & Reker) wants scientific advice to be a “fixed boundary condition” and calls science-grounded management “undoubtedly a more objective way” (p. 422). The main chapter shows that the official science was badly wrong for years (1986–89), and argues that authority “built on a certain control of science” is itself part of the problem (p. 419). Both can be reconciled, as “better, plural, reviewed science and more deference to it”, but the chapter never does so explicitly. The Canadian record shows two distinct failure modes in sequence: first science failed (overconfident assessment), then politics failed (overriding corrected advice). The lessons differ for each.

Advocacy in Panel 17.2.

• The panel says stocks “continued to worsen” (p. 421), yet its own figures show NE Atlantic improvement (94% to 63% overfished; the MSY share rising in Figure 17.1).
• The economic case relies on a single estimate from an advocacy-oriented think tank (NEF; my characterisation), with a units typo.
• The panel takes a position on a pending reform. It welcomes the “ambitious measures” of the 2011 Commission proposal but says it is “questionable” whether the system will deliver unless advice is respected (p. 422).
• It presents political bargaining over TACs chiefly as “narrow political interests and negotiations”, “short-term economic interests” and a “vested political decision-making system” (p. 422). It does not engage the argument that socio-economic considerations are a legitimate part of democratic decisions, the position set out on the Minister’s side in the Nunavut case (p. 420).
• It says there is “sufficient available knowledge” for an ecosystem approach (p. 422), while also stressing that data are missing for about two-thirds of NE Atlantic TAC stocks (p. 421).

Counter-arguments not engaged.

• The social cost of precautionary closures to communities: Norway’s 40% income drop is noted but not weighed.
• The possibility that local knowledge can also be wrong or self-interested. Fishers’ data reliability is doubted in the panel itself (p. 412), but that doubt is not carried into the lessons.
• The practical difficulty of “diversity-oriented” management at the scale of hundreds of stocks.
• Whether more complex models would actually have forecast the collapse.

Fair in the other direction. The chapter:

• Does acknowledge the lack of a consensus replacement for simple models (p. 415).
• Acknowledges the low pedigree of causal knowledge (Table 17.1).
• Concedes that diversity-oriented analyses “tend not to provide insights into causal generalisations” (p. 409).
• Admits that the effects of TAC–advice mismatches “remain uncertain” (p. 421).
• In Panel 17.2, lists “the lack of scientific knowledge and ineffective management options” alongside decision-making as causes of EU failure (p. 421).
• Treats IPBES’s recognition of local knowledge as an ongoing struggle rather than a success (p. 416).

The central Canadian narrative is well corroborated by official reviews that are independent of the authors: Alverson, Harris and the Auditor General.

Notable quotes#

1. “Our systemic failure to anticipate such surprises forms the core of this chapter.” (p. 407)
2. “The Norwegians had understood that no evidence of harm is not the same as evidence of no harm.” (p. 411)
3. “The surprise was that they had used the same data and come out with startlingly different conclusions.” (Panel 17.1, McGlade; p. 412)
4. “Thus catch rates might continue to increase even as the stock was collapsing. Unfortunately, DFO accepted the data as they stood.” (Panel 17.1; p. 413)
5. Scientists “‘lulled by false data signals and, to some extent, overconfident of the validity of their predictions’” (Leslie Harris 1990, quoted in Panel 17.1; p. 413)
6. Ministers rejected the cut, “supporting their decision by saying that the scientists had been wrong before.” (Panel 17.1; p. 413)
7. “‘The credibility of DFO scientific advice was not questioned. The northern cod stock continues to increase, but perhaps not as fast as projected several years ago’.” (DFO Director General, Fo’c’sle, Feb 1988; p. 414)
8. “…many assessment scientists had little contact with the fishing industry or actual fish, dealing rather with ‘paper fish’.” (p. 415)
9. “…it will take more than the collapse of a resource or a shift in an ecosystem to displace the existing form of authority that has been built on a certain control of science.” (p. 419)
10. “The assumption has been that all interactions can be adequately understood and that all future states are contained within the present structure of the socio‑ecological system.” (p. 424)

Open questions#

1. Norway. - What exactly did Norway do in 1989–90? A full cod moratorium, or drastic TAC cuts, vessel quotas and early closures? - Did Norway ban fishing on spawning grounds, or only restrict trawling there, given the Lofoten spawning fishery? - Who was Fisheries Minister when the measures were taken, and when did Pettersen make the quoted remark? - Primary sources would settle this: ICES advice for 1989–90, Norwegian Ministry records, and the Joint Norwegian–Russian Fisheries Commission TACs.
2. Nunavut judgment. What does Nunavut Tunngavik Inc. v. Canada ([1998] 4 F.C. 405) actually say? Is the “absolute discretion… numerous fluctuating policy concerns” passage the Minister’s argument, or the court’s paraphrase of statute (Fisheries Act s. 7) and Supreme Court authority (Comeau’s Sea Foods, 1997)? Is the “academy of science” line quoted from earlier case law? On what ground was the decision unlawful (background suggests the Nunavut land-claims agreement)? All background; unverified. Worth retrieving; the government’s FJA reports site was unavailable when I tried.
3. Primary Canadian sources. Would the Harris 1990 review, the Alverson 1987 report, the 1986 Keats et al. review and the Auditor General’s 1997 chapters 14–15 support the chapter’s reading, including the “tuning” dynamic and DFO’s admission that it had evidence in 1986? Andrew, these would be worth retrieving if you can.
4. Publication decline. Where does the claim that papers on northern cod fell from more than 10 a year to 2 come from, and over which years? Does the career-incentive explanation hold? (The chapter’s own logic is also puzzling: a system that rewards peer-reviewed publication would, on its face, encourage papers rather than discourage them. It presumably means assessment work and basic biology were not rewarded, but it does not say so.)
5. Science versus politics. How much of the Canadian failure was science failure and how much political override, and does the distinction change the lesson? The chapter runs them together.
6. Better models. Would a “diversity-oriented” approach actually have produced an earlier or clearer warning in 1986? Or was the decisive factor willingness to credit external re-analysis and fishers’ observations, which is institutional rather than methodological?
7. Science’s authority after the collapse. Is the claim that science-based authority survived the collapse with only “modest changes” (p. 419) borne out by post-1992 DFO reforms, changes to CAFSAC and the advisory structure, and later assessment practice?
8. Hindsight pointers. These are outside the report and from my background knowledge, unverified. They belong to the separate hindsight strand. - Northern cod status after 2013, including a 2024 Canadian decision to end the moratorium after a revised assessment model. The model-dependence of stock status is itself an echo of this chapter. - Barents Sea cod after 2013: record-high biomass around then, followed by declines and sharp TAC cuts in the mid-2020s. This tests the Norway-as-model narrative. - Outcomes of the 2013 CFP reform: the MSY-by-2015/2020 objective, the landing obligation, and the TAC–advice gap in the NE Atlantic versus the Mediterranean. - Planetary boundaries: later assessments claiming several boundaries crossed, against the chapter’s “many decades of tantalising ignorance”. - Tipping-element research since 2013. - The generic early-warning-signal literature, set against the chapter’s scepticism of generic threshold models. - How IPBES went on to operationalise indigenous and local knowledge. - The outcome of the Anthropocene epoch proposal.
9. Missing material. What was cut? The reference list includes the Challenger commission (Rogers 1986) and Ferguson’s Civilization, neither of which is cited. Was there a longer version of the chapter with a Challenger or organisational-failure comparison?

Audit log#

Independent audit against the full text extract (all 22 pages) and the PDF (pp. 407–428; front matter pp. 5 and 8; Annex 2 p. 703; index; Figure 17.1 re-read from a page render).

• Authors: confirmed McGlade signs the Preface as Executive Director (p. 8), van den Hove is on the editorial team, Belchior and Reker are contributing authors, and Funtowicz and Ravetz are on the advisory board (p. 5); no change needed.
• Authors: Stokes, McGlade & Law 1993 marked as reference-list only (not cited in text); “began her career” softened to “in 1980 she was newly recruited”.
• Panels table: “Heroic narrative” changed to “Celebratory narrative”; Box 17.1 “hagiographic” changed to “celebratory, unsourced”; added the “Scripps Institute” slip.
• Panels table: Panel 17.2 “backs the 2012 CFP reform” corrected; the panel welcomes the proposal but doubts it will work without respect for advice.
• Other-side voices: added Crosbie’s three-factor explanation, and noted that the Norwegian minister is quoted with approval, in reported speech only.
• Abstract: “Yaffee” is not named in the abstract (only at p. 415); added “competition replaces co-operation”.
• 17.1: added the Carson epigraph (omission).
• 17.2: added the “inverse relationship between the number of cases and variables” framing (omission).
• 17.2.1: completed the “works best when…” quote; added the chapter’s own concession that diversity-oriented analyses “tend not to provide insights into causal generalisations” (important omission).
• 17.2.2: marked the MSY “assumptions carried across domains” line as my gloss, not the chapter’s.
• 17.2.2 dating: added that the panel itself says “collapse… in the 1980s” (p. 412), and quoted Annex 2’s “1992 The stock collapses” verbatim (checked in PDF).
• 17.2.2: Pettersen’s words are reported speech in the chapter, not a direct quotation; the note is corrected, as are the downstream references (“reported admission”, “ministerial remark”).
• 17.2.2: “in the same section” corrected to “on the same page (17.2.3)” for “myth-like”.
• Panel 17.1: Memorial review requester corrected to “the industry” (panel wording), with Annex 2’s “inshore fishers” given separately.
• Panel 17.1: Templeman’s mechanism is presented in the panel as “one potential explanation”, not a documented fact; this is now flagged.
• Panel 17.1: completed the Parsons quote and flagged its “early 1980s” (as printed, verified in PDF) as a probable slip.
• Panel 17.1: added the undated (“Meanwhile”) lawsuit and DFO’s maintained “no evidence” line.
• Panel 17.1: corrected a misattribution. “Evidence as early as 1986” is the panel’s statement, not clearly the Auditor General’s, and the AG model quote covers “most of the key groundfish stocks”.
• Panel 17.1 Barents: marked the recovery-by-prompt-action attribution as the author’s untested claim; added “in the end” for the EUR 50 million.
• 17.2.3: restored the hedge “may have contributed to the overestimation of biomass” on the single-stock-unit flaw.
• Table 17.1: added a background flag that NUSAP level-2 peer acceptance is “Medium” (possible slip).
• 17.3.1: the Anthropocene quote is made fuller; noted that the chapter’s own Zalasiewicz et al. 2011 source title ends in a question mark.
• Table 17.2: added the Gorongosa “war of Independence in the 1980s” error (background).
• 17.4.1: softened the Everglades “conflicts with Table 17.2” to “sits awkwardly with” (the table may describe the earlier phase).
• 17.4.1 Nunavut: added background that the “absolute discretion” wording is Fisheries Act s. 7 language, and that the “academy of science” line appears to quote earlier case law, so it may be the court’s statement of law (unverified).
• 17.4: added the chapter’s framing questions about who gains and loses power in a crisis.
• Panel 17.2: added the panel’s concession that “lack of scientific knowledge and ineffective management options” also contribute (important omission).
• Panel 17.2: added “impacts… remain uncertain” and the OSPAR qualifier (Greater North Sea, low natural disturbance); set the Habitats Directive 2%/10% in the panel’s marine-biodiversity context.
• Panel 17.2: added the MSFD ecosystem-approach objective, the “sufficient available knowledge” claim, the EP 2012 recognition, and “vested political decision-making system”.
• Figure 17.1: re-read from the PDF render. Corrected the “no advice” range (about 31–42%), added the “advice other than MSY” category and the 2006 dip, and noted that overfished plus MSY stays about 35–38%.
• Table 17.3: added the Demographic row’s outcome and full responses.
• References: EU Court of Auditors 2011 removed from “uncited” (it is probably the “EU, 2011” cited in Panel 17.2); added May 1982 and Holling 1973/1996 as cited but missing from the list; added the Schellnhuber 1999/1994 and “Villasante et al.” slips.
• Timeline A: the Memorial row, the 1986 “evidence” row and the 1997 AG row are corrected for attribution and scope; the lawsuit is marked undated; added Annex 2’s “halve the offshore catch”.
• Lags: corrected “TACs meanwhile held or raised against advice”. The TAC rise was for 1988, before the December 1988 advice; afterwards the cut was rejected and the 1990 reductions were insufficient.
• Harms: corrected the misreading that “the chapter anticipates” job losses (that was the ministers’ feared cost of a quota cut); the CAD 2.56 billion is now dated to about 1989.
• Who bore the costs: tightened to the chapter’s wording (“social structure of the fishery”), removing the claim that it was specifically the inshore fishers’ social structure.
• Timeline B: spawning-ground ban moved to undated (the chapter gives no date); lag measured from the autumn 1989 survey.
• Mechanisms: re-analysis bullet softened (“changing assumptions rather than new data” is marked as inference; Alverson added fishers’ evidence); hyperstability bullet made conditional; Crosbie’s inclusion of stock overestimation noted; credibility-trap inference labelled.
• Economics: removed “patronage” (not in the source); “costs offset by recovery” changed to the chapter’s untested attribution; Nunavut discretion cross-referenced to the attribution caveat.
• Mental models: “central explanatory variable” softened to one of the chapter’s main contrasts.
• Lock-in: “infrastructure lock-in” labelled as my framing, and the Table 17.2 wording is quoted.
• Insight 1: strength downgraded from strong to moderate–strong; “confirmed by Harris and the AG” corrected to “consistent with”.
• Insight 4: removed the unsupported “cheap” and “under different assumptions”; reframed as outside re-analysis.
• Insight 9: “systematically” changed to “tends to”, and strength changed from moderate to suggestive–moderate (it rests mainly on McGlade’s testimony).
• Insight 14: “least powerful” changed to “resource-dependent community”, in line with the chapter’s wording.
• Limitations: “written by” changed to “co-written by”; added new flagged errors (Parsons date, AG attribution, Gorongosa, VPA derivation) and the Zalasiewicz question mark.
• Panel 17.2 advocacy: corrected “supports reform” to welcome-with-doubts; “only” changed to “chiefly”, with the panel’s own wording; added the “sufficient knowledge” versus data-gap tension.
• Fair in the other direction: added the p. 409 concession and Panel 17.2’s acknowledgement of knowledge gaps.
• Open questions: Q2 expanded with the statute and case-law leads; Q4 notes the chapter’s puzzling publication-incentive logic.
• Digest: fixed the AG attribution; added Panel 17.2’s concessions; made the hyperstability mechanism conditional; downgraded insight 1; reworded insight 4 and qualified insight 9; “co-written”; expanded the caveats (dating, Parsons, Nunavut attribution, reported speech, the chapter’s own concessions, the “sufficient knowledge” tension); flagged the Norway recovery attribution.
• Strand check: no contemporary technologies or companies outside the source are mentioned in the notes or digest; “quants” and Rumsfeld appear in the source itself.