MAC Lab #001 — Minimum Viable Consciousness (Requirements)

Date: Thursday, July 23, 2026
Time: 6:30–8:30 PM
Format: In-person workshop — max 20 participants, registered MAC members
Facilitator: Loki
Location: Downtown Vancouver (details via MAC Lab registration)

“True progress comes from making theories vulnerable to falsification, not protecting them.”
— Lucia Melloni, Cogitate Consortium / Max Planck Institute for Empirical Aesthetics[^1]

What Is This Lab About?

There are now more than 325 distinct theories of consciousness. They range from well-established neuroscientific frameworks like Integrated Information Theory (IIT) and Global Workspace Theory (GWT) to quantum, panpsychist, enactivist, and idealist accounts. Most have accumulated their own bodies of supportive evidence — yet research shows this support can be predicted from methodological choices alone, independent of actual findings. The theories rarely meet head-on.[^2][3][^4][5][^6]

The field is experiencing a methodological crisis: too many theories, too little cross-talk, and almost no shared criteria for what it would even mean for one theory to be better than another.[^7][8]

:AB #001 addresses this directly.

Instead of debating the theories themselves, participants will design an adversarial filter — a structured set of challenges that any viable theory of consciousness should be able to answer. The lab draws on a real and growing tradition in consciousness science: adversarial collaboration, the practice of bringing competing camps together to jointly construct tests neither side can easily dismiss.[^9][10]

The Proposed Research Question

Can a collaboratively designed adversarial challenge-set reliably distinguish stronger from weaker theories of consciousness — and if so, what is the minimum such a set requires?

This is an experimental question, not merely a philosophical one. It targets the process of theory evaluation, not any specific theory. The group will aim to answer it by designing and piloting a small prototype filter.

Assertion:  A viable theory must provide a coherent explanation or mechanism for a given challenge.
A strong theory provides effective, mechanistically grounded answers across most or all challenges.

Background: Why Adversarial Filtering?

The scientific case for adversarial approaches to consciousness has been building steadily.

In 2021, the ConTraSt database — a systematic review of 412 consciousness experiments — found that whether a study supports a given theory can be predicted from the study’s methodological choices, regardless of results. This is a profound methodological indictment: theories are not being genuinely tested; they are being confirmed.[^11][3]

In 2025, the Cogitate Consortium — an international team including proponents of both IIT and GWT — published a landmark study in Nature that directly and adversarially tested both theories using preregistered predictions, 256 participants, and three neuroimaging modalities. Both theories were substantially challenged. Neither “won.” But the approach — borrowed from Nobel laureate Daniel Kahneman’s concept of adversarial collaboration — was described as a model for how science should advance in the face of deadlock.[^12][13][^14][10][^15]

Our lab scales this spirit down to a advance preparation, a 2-hour collaborative session, and a following deliverable – replacing lab equipment with carefully constructed conceptual challenges.

Separately, philosophers have proposed formal criteria for what consciousness theories must achieve. Doerig, Schurger, and Herzog (2020) argued for theory-neutral “hard criteria” — benchmarks empirical theories should meet regardless of their theoretical commitments. Others have proposed Bayesian frameworks for quantifying empirical support across theories. All point toward the same conclusion: the field needs shared, reusable evaluation tools — the kind this lab begins to build.[^16][17][^18][19]

Exemplar Challenge

What would a challenge look like? See below for a model for how a well-formed challenge looks and functions:

The Synesthesia Challenge
“Under this theory, provide a mechanism or explanation for the phenomenon of synesthesia — the reliable, involuntary co-activation of sensory modalities (e.g., hearing a sound that reliably evokes a specific colour).”

This challenge is effective because:

• It targets a well-documented, studied, non-trivial phenomenon that is not a dismissable edge case
• It requires a mechanistic account, not just a verbal acknowledgement
• It differentiates theories: IIT can invoke changes in integration architecture to account for altered experience; GWT must explain how a stimulus activates cross-modal workspace broadcasts that are not typically triggered; Higher-Order Theories must address whether the co-activation involves second-order representations[^20][21]
• It is answerably bounded — a theory can reasonably attempt it without requiring new empirical work

The Synesthesia Challenge is not exotic. MAC’s own Fiann O’Hagan has presented synesthesia — alongside Capgras syndrome — as precisely the kind of boundary case that reveals the internal structure of a theory of consciousness.[^2]

Approaches to the Goal of an Adversarial Filter

Four approaches are worth considering. Participants the others prior to committing to the lab’s primary method.

Approach	Core Logic	Strength	Limitation
1. Phenomena Coverage	List canonical conscious phenomena; evaluate theories by how many they address	Broad, systematic	Tends to reward loose verbal fits
2. Mechanism Constraint	Theories must provide mechanistic (not just descriptive) accounts	Rigorous, anti-vague	Hard to apply consistently across theories
3. Predictive Power	Assess novel, testable predictions each theory makes	Highest scientific value	Requires empirical follow-up
4. Adversarial Challenge (Selected)	Design targeted challenges that expose theoretical gaps	Collaborative, portable, fast	Depends on quality of challenges designed

 

The lab adopts Approach 4. It is the most tractable in a 2-hour group setting, produces a reusable artifact, and maps most directly onto the adversarial collaboration tradition now gaining traction in professional consciousness science.[^22][23]

Background Reading

Estimated preparation time: 5–10 hours total. Not all items need to be read in full — annotations and abstracts are fine for some.

On the Problem of Too Many Theories

• Kuhn, R.L. (2024) — “A Landscape of Consciousness: Toward a Taxonomy of Explanations and Implications.” Covers 325+ theories along a physicalist-to-idealist spectrum. [Consciousness Atlas interactive version at com][^4][6]
• Yaron et al. (2022) — “The ConTraSt Database for Analysing and Comparing Empirical Studies of Consciousness Theories.” Nature Human Behaviour. Shows how methodological choices — not results — predict which theory a study will “support.”[^3][24]
• Kirkeby-Hinrup et al. (2025) — “Methodological Issues in Consciousness Research: Theory Comparison, the Role of Empirical Evidence, and a Replication Crisis.” Frontiers in Psychology. — Accessible overview of why the field is stuck and what rigorous comparison requires.[^8][7]

On Adversarial Approaches

• Cogitate Consortium / Melloni et al. (2025) — “Adversarial Testing of Global Neuronal Workspace and Integrated Information Theories of Consciousness.” — The landmark study. Read the abstract and the discussion section closely.[^13][12]
• Kahneman, D. (via Templeton) — “Adversarial Collaboration” (short video, ~15 min). — The conceptual origin of adversarial collaboration, from the Nobel laureate who developed it.[^9]
• Wikipedia: Adversarial Collaboration — Clean overview of the method and its history.[^10]
• Reflections on Adversarial Collaboration from the Adversaries (2025) — PMC/PubMed. Survey of 29 scholars across 13 adversarial collaboration projects. Most common outcome: not a “winner,” but deeper understanding.[^15]

On Evaluation Criteria

• Doerig, Schurger & Herzog (2020) — “Hard Criteria for Empirical Theories of Consciousness.” Cognitive Neuroscience. — Foundational paper proposing theory-neutral evaluation standards.[^17][25]
• Chis-Ciure (2024) — “Quantifying Empirical Support for Theories of Consciousness.” Frontiers in Psychology. — Proposes a Bayesian methodology for comparing support across theories.[^18][19]
• Signorelli, Szczotka & Prentner (2021) — “Explanatory Profiles of Models of Consciousness.” Systematic classification framework.[^26][27]

On Core Theories (Choose 1–2 to prepare deeply)

• IIT: Tononi et al. — Overview at Scholarpedia; Wikipedia[^28][29]
• GWT/GNWT: Baars, Dehaene — Wikipedia overview; Mashour et al. (2020) Neuron review[^21][30]
• Predictive Processing / Free Energy: Friston — accessible summaries widely available
• Higher-Order Theories: Lau & Rosenthal — SEP and Wikipedia entries

On Evidence for Challenge Cases

• Blindsight — BBC Future overview; PMC neuroscience review[^31][32][^33]
• Split-brain — Brain journal: “Split Brain: Divided Perception but Undivided Consciousness”[^34]
• Synesthesia — Ramachandran & Hubbard (2001), Philosophical Transactions of the Royal Society B (widely available)

What Makes a Strong Challenge?

The quality of the filter depends entirely on the quality of the challenges.

Before contributing a challenge, apply these criteria:

Criterion	Question to Ask
Relevance	Does this target a central feature of consciousness, not a peripheral curiosity?
Non-triviality	Could any vague claim pass this challenge, or must a theory actually work for it?
Discriminative power	Will different theories give meaningfully different answers?
Mechanistic demand	Does it require a mechanism, not just a description or relabelling?
Empirical grounding	Is the phenomenon documented and agreed-upon, not just hypothetical?
Answerability	Could a reasonable theory, in principle, attempt an answer?
Scope specificity	Is the challenge narrow enough to evaluate, or too broad to fail?

 

A weak challenge looks like: “Under this theory, is there an explanation for redness.”
A strong challenge looks like: “Under this theory, what accounts for the fact that some split-brain patients report a unified sense of self despite functionally independent hemispheres?”[^34]

Challenge types to consider:

• Dissociation challenges — phenomena where perception and awareness come apart (blindsight, blindsight vs. sight, pain asymbolia)[^33][31]
• Cross-modal challenges — phenomena involving unexpected sensory binding (synesthesia, the McGurk effect)
• Unity/boundary challenges — phenomena testing what counts as one experience (split-brain, binocular rivalry)
• Gradation challenges — phenomena involving degrees of consciousness (anaesthesia, sleep stages, psychedelics)
• Absence challenges — what a theory predicts for systems that are clearly not conscious (a thermostat, a simple reflex arc)
• Emergence challenges — when and how does a physical system become conscious under this theory?

Lab Activities

Before the Lab (~5–10 hours per person)

1. Read — Select at minimum: the ConTraSt database paper, the Cogitate Consortium Nature paper abstract + discussion, and the Doerig hard criteria paper. Skim one or two theory overviews.[^13][17][^3]
2. Prepare one challenge — Draft 1–2 candidate challenges using the criteria above. Write them in this form:

“Under this theory, provide a mechanism or explanation for [phenomenon].”
Append a brief note (2–3 sentences) explaining why you think it is discriminative.

3. Optional: apply your challenge — Pick one theory and sketch how it would respond. This preparation enriches the lab discussion considerably.
4. Submit your challenge to the facilitator at least 48 hours before the lab (exact submission method via MAC mailing list).

Facilitator note: Collect submitted challenges in advance. Cluster them thematically (e.g., dissociation, unity, gradation) and prepare a short visual summary (whiteboard or slide) showing the distribution. This helps the group make a conscious selection rather than simply voting by familiarity.

At the Lab (2 hours)

• Review submitted challenges
• Select top 5 challenges
• Run a preliminary analysis with 3 selected theories

After the Lab (TBD)

• Refine challenges — Volunteers edit and sharpen the selected set based on lab discussion
• Apply the protocol — Individuals or pairs apply the full challenge set to one theory and write up a short response profile (1–2 pages)
• Develop scoring — Working group drafts one or more scoring frameworks (see below) for the group to evaluate
• Synthesis — Facilitator or volunteer compiles response profiles into a comparative document

Assessment Protocol

The following protocol can be applied to any theory after the lab:

1. Interpretation Phase — Clarify the theory’s key terms and scope. What does it define as “consciousness”? What is explicitly out of scope?
2. Challenge Response Phase — For each challenge, articulate the theory’s mechanism or explanation. Aim for 100–200 words per response. Use the theory’s own vocabulary.
3. Evaluation Phase — Score each response (see Scoring Methodologies below) along agreed dimensions.
4. Evasion Audit — Flag any response that redefines the phenomenon rather than explaining it, or invokes mechanisms the theory does not independently posit.
5. Cross-comparison Phase — Compare response profiles across theories. Where do they diverge? Where do they converge on the same answer despite different frameworks?

Proposed Outcomes

The lab is intended to generate:

• A prototype challenge set (5–10 challenges) with documented selection rationale
• A draft assessment protocol describing how to apply the set to any theory
• An initial scoring framework chosen or composed by the group
• Optional: a comparative response profile for 2–3 theories, produced post-lab

Longer-horizon outputs the group may consider:

• A short working paper or report — suitable for distribution or submission
• A web-accessible tool where any theory can be evaluated against the challenge set
• A challenge library — an expanding open resource for the broader consciousness research community
• Submission to a collaborative annotation project building on existing databases like ConTraSt[^3]

MAC Lab #001 is part of a new workshop series complementing the monthly MAC Deepdives. Labs emphasise design, experiment, and output over discussion alone.  Pre-registration required via the MAC mailing list. Previous MAC Deepdive attendance preferred.
Questions: [email protected]

References

1. Michael Pitts and Cogitate Consortium Publish Study on …
2. Vancouver AI Community Meetup MAC Takeover – 2/26 – YouTube – Mind, AI & Consciousness takes over Vancouver AI for a special February 2026 edition exploring what …
3. The ConTraSt database for analysing and comparing empirical … – We provide a bird’s eye view of studies that interpreted their findings in light of at least one of …
4. A landscape of consciousness: Toward a taxonomy of explanations … – Diverse explanations or theories of consciousness are arrayed on a roughly physicalist-to-nonphysica…
5. A Landscape of Consciousness: Toward a taxonomy of explanations … – Read about diverse theories of consciousness from materialist/physicalist to nonmaterialist/nonphysi…
6. Consciousness Atlas: Interactive Visualization of 325+ Theories from … – The Consciousness Atlas is a straightforward visualization tool: it presents Kuhn’s taxonomy as a si…
7. theory comparison, the role of empirical evidence, and a replication … – That is: collect evidence for each theory, compile the sets of evidence, validate the claimed empiri…
8. theory comparison, the role of empirical evidence, and a replication … – Ellia F., Chis-Ciure R. (2022) … Quantifying empirical support for theories of consciousness: a te…
9. Adversarial Collaboration featuring Lucia Melloni,… – Adversarial collaboration is an innovative approach that may help advance science, despite fragmenta…
10. Adversarial collaboration – Wikipedia – Adversarial collaboration is a modality of collaboration wherein opposing views work together in ord…
11. The ConTraSt database for analysing and comparing empirical … – We provide a bird’s eye view of studies that interpreted their findings in light of at least one of …
12. Rethinking Consciousness: When Science Puts Itself to the Test – International Consortium Conducts First Direct, Adversarial Test of Two Leading Theories of Consciou…
13. Adversarial testing of global neuronal workspace and integrated … – Different theories explain how subjective experience arises from brain activity. These theories have…
14. Make science more collegial: why the time for ‘adversarial … – Nature – Bringing together proponents of rival theories to test their ideas against each other can advance sc…
15. Reflections on adversarial collaboration from the adversaries – PMC – Adversarial collaborations offer the promise of breaking deadlocked debates, resolving disputes, and…
16. Criteria for empirical theories of consciousness should focus on the … – Doerig and colleagues put forward the notion that we need hard and theory-neutral criteria by which …
17. Hard criteria for empirical theories of consciousness – PubMed – First, we argue that consciousness is a well-defined topic from an empirical point of view and motiv…
18. Quantifying empirical support for theories of consciousness – PubMed – Understanding consciousness is central to understanding human nature. We have competing theories of …
19. Quantifying empirical support for theories of consciousness – Frontiers – A methodology to quantify the divergent sets of empirical support proposed in favor of extant theori…
20. Does Integrated Information Theory (IIT) feel too woo-woo to anyone? – In synesthesia, the same input produces different sentient experiences when integration changes. In …
21. Global workspace theory – Wikipedia
22. Accelerating Research on Consciousness: An Adversarial… – In the context of an adversarial collaboration, it will test two prominent theories of consciousness…
23. No Clear Winner: 7-Year Brain Study Tests Rival Theories of …
24. The ConTraSt database for analysing and comparing … – PubMed – We provide a bird’s eye view of studies that interpreted their findings in light of at least one of …
25. Hard criteria for empirical theories of consciousness
26. Theories of consciousness and psychiatric disorders – Signorelli et al., 2021. C.M. Signorelli, J. Szczotka, R. Prentner. Explanatory profiles of models o…
27. [PDF] Explanatory profiles of models of consciousness – towards a … – Quantifying empirical support for theories of consciousness: a tentative methodological framework …..
28. Integrated information theory – Scholarpedia
29. Integrated information theory – Wikipedia – IIT aims to explain which physical systems are conscious, to what degree, and in what way. The theor…
30. Microsoft Word – FINAL TEXT V4, GNW Review, Neuron, 1-14-20.docx
31. Blindsight: the strangest form of consciousness – BBC – Some people who have lost their vision find a “second sight” taking over their eyes – an uncanny, su…
32. The nature of blindsight: implications for current theories of … – PMC – Blindsight regroups the different manifestations of preserved discriminatory visual capacities follo…
33. The nature of blindsight: implications for current theories … – Blindsight regroups the different manifestations of preserved discriminatory visual capacities follo…
34. Split brain: divided perception but undivided consciousness – The canonical view of split-brain patients is that splitting the brain also splits consciousness, wh…
35. 18 OCTOBER 2019 • VOL 366 ISSUE 6463 293