Narrative Self Café v23a: Minding Over What Matters
Posted: October 21, 2026 Filed under: Centroids | Tags: agency, consciousness, empathy, identity, philosophy Leave a commentWrite an evolutionary interlude following v23: Being Conscious (Of), returning to CASH, BASH, HASH, RIF, and the Bodily Affect Divide.
ChatGPT Prompt
Prologue: The Question We Kept Asking
DAMASIO: Consciousness begins with feeling.
FRISTON: And feeling begins with regulation.
JÉKELY: And regulation did not begin with nervous systems.
VARELA: Nor did nervous systems begin with representation.
BARRETT: Then perhaps we should begin somewhere else.
DAMASIO: Where?
BARRETT: With the question an organism must answer before it can know what anything means.
FRISTON: Which question?
BARRETT: What is wrong?
A pause.
JÉKELY: That is already too sophisticated.
BARRETT: Why?
JÉKELY: Because it presumes an organism capable of asking what.
VARELA: And perhaps one capable of distinguishing wrong from merely different.
DAMASIO: But that is precisely what homeostasis provides.
FRISTON: Not quite. Homeostasis provides regulatory dynamics. It does not necessarily provide an explanation of their disturbance.
BARRETT: Then let us separate the questions.
She writes on the café blackboard.
- HOW MUCH?
- WHERE?
- WHAT CAUSED IT?
- WHAT WOULD CHANGE IT?
BARRETT: Perhaps we have been calling all four of these things feeling.
DAMASIO: I certainly have not.
BARRETT: No. But our explanations have a habit of collapsing them.
VARELA: Then we should reconstruct the distinctions.
JÉKELY: From the beginning?
FRISTON: From before the beginning, if necessary.
Act I — The Body Was Already Moving
JÉKELY: Our earlier mistake was to imagine that nervous systems evolved because organisms needed to process information about the world.
FRISTON: That is hardly a mistake. Organisms do need information.
JÉKELY: Of course. But the body was already doing something before it needed to know what it was doing.
He draws a pulsing oval.
JÉKELY: Excitable tissue. Recurrent contraction. Mechanical deformation. Recovery. Another contraction.
VARELA: A system sustaining its own activity across time.
JÉKELY: Precisely. Movement does not require a map. Rhythm does not require representation.
FRISTON: But it does require coordination.
JÉKELY: And that is the evolutionary opening.
RIF — Recurrent Information Field
JÉKELY: A recurrent information field is not necessarily a brain, or a mind, or even a living thing. It is an organized dynamical regime whose present activity constrains its future activity.
VARELA: Endogenous activity is primary. External inputs perturb it.
BARRETT: So the organism does not begin as a passive receiver waiting for the world to deliver information.
JÉKELY: Exactly. It begins already in motion.
DAMASIO: But movement alone is not enough. A body must remain viable.
FRISTON: Which brings us back to regulation.
HASH — Hierarchical Adaptive System Hysteresis
FRISTON: History reshapes the available transitions. Regulatory constraints operate across nested levels. Some trajectories become easier; others become harder.
VARELA: One regulates. One hums.
BARRETT: HASH and RIF.
VARELA: Different descriptions of organization, not competing substances.
JÉKELY: And neither yet requires an organism to represent the source of a disturbance.
DAMASIO: But as movement becomes more complex, that source begins to matter.
JÉKELY: Yes. Which is why I would rather ask what nervous systems had to stabilize before asking what they had to know.
Act II — The Twitch That Knew Nothing
BARRETT: Consider an infant animal twitching in its sleep.
FRISTON: Dreaming?
BARRETT: Perhaps. But let us leave subjective experience out of it.
DAMASIO: A rare act of restraint in a discussion of consciousness.
BARRETT: The twitch is a motor event. It produces a change in bodily configuration. That change generates sensory feedback.
JÉKELY: An endogenous movement encountering its own mechanical consequences.
VARELA: A recurrent system becoming coupled to its body.
BARRETT: Exactly.
Research on sleep-related twitching in developing animals suggests that such movements contribute to the organization of sensorimotor circuits. They may help calibrate the relationship between motor activity and bodily feedback.
This does not establish that dreaming itself exists for mechanical calibration. It gives us a narrower and more useful observation: internally generated movement can produce structured sensory consequences that developing systems can learn from.
FRISTON: So we have a calibration loop.
Motor activity → Bodily deformation → Proprioceptive feedback → Modified motor dynamics
JÉKELY: Notice what is absent.
DAMASIO: A goal?
JÉKELY: A representation. The animal need not think, I moved my left leg.
VARELA: Nor must it possess a conceptual distinction between itself and the world.
BARRETT: Yet something important can be learned.
FRISTON: Which motor dynamics reliably produce which mechanical consequences.
BARRETT: And perhaps, eventually, which sensory changes are consequences of endogenous activity rather than independent environmental events.
DAMASIO: A functional precursor to the distinction between self and nonself.
VARELA: A precursor, perhaps. Not yet the conscious ownership of experience.
DAMASIO: Agreed.
JÉKELY: Now consider what happens when the organism moves through a heterogeneous environment.
Movement → Environmental encounter → Sensory perturbation → Modified movement
FRISTON: And a third:
Environmental encounter → Physiological consequence → Regulatory adjustment
BARRETT: So we have at least three distinguishable relationships: What does my motor activity do to my body? What does my bodily movement do to my encounter with the environment? What does that encounter do to my physiological condition?
VARELA: The crucial development may not be a new information-processing module.
JÉKELY: It may be the differentiation and calibration of couplings among dynamics that already exist.
FRISTON: And the coupling strengths themselves can change with history.
BARRETT: CASH again.
FRISTON: Trajectories deform the landscape. The landscape constrains future trajectories.
Act III — Three HASHes and a Nematode
A nematode appears on the tablecloth. Nobody asks how.
DAMASIO: A convenient organism.
JÉKELY: Convenient precisely because we need not settle whether it has subjective experiences.
BARRETT: Or whether it is asking philosophical questions.
FRISTON: Its behavior still presents a genuine problem.
INTEROCEPTION · EXTEROCEPTION · MOBILITY
FRISTON: Suppose we model the nematode as three coupled HASH organizations. Interoception tracks physiologically relevant internal conditions. Exteroception is sensitive to environmental gradients and perturbations. Mobility generates the ongoing locomotor dynamics through which the organism encounters its surroundings.
JÉKELY: And proprioception?
FRISTON: That is the interesting complication.
VARELA: It might be a distinct adaptive organization. Or it might be the feedback structure through which mobility calibrates itself.
BARRETT: We should not decide that by drawing another circle.
JÉKELY: Quite. Anatomical distinctions, dynamical distinctions, and analytical distinctions are not necessarily the same.
DAMASIO: But the functional question remains.
FRISTON: How do these systems constrain one another?
BARRETT: And how does the organism discover which environmental distinctions matter for its internal condition?
JÉKELY: A nematode following a chemical gradient need not represent food as food. Its recurrent locomotor dynamics can be biased by sensory and physiological conditions.
FRISTON: Yet that bias can depend on history. The same sensory stimulus can produce different behavior depending on the organism’s state and previous encounters.
VARELA: That is precisely why a simple stimulus-response account is inadequate.
DAMASIO: But what makes the stimulus important?
FRISTON: Its consequences for viability.
BARRETT: And there is the first question we can now ask more precisely.
How does a physiological constraint become a change in the organization of attention and action?
JÉKELY: Careful. You have introduced attention before showing that the nematode possesses anything resembling human attention.
BARRETT: Then call it selective allocation of sensory and behavioral capacity.
VARELA: An operational description. Not a claim about conscious awareness.
DAMASIO: And not necessarily a single centralized mechanism.
BARRETT: Agreed. But the allocation problem exists even when the vocabulary is provisional.
Act IV — The Boundary That Did Not Break
FRISTON: We should revisit BASH.
BASH — Bottlenecked Adaptive System Hysteresis
FRISTON: A complex adaptive system interacts with other systems through constrained interfaces. The internal dynamics may possess many more degrees of freedom than can be expressed across a particular boundary.
VARELA: So the boundary becomes a bottleneck.
FRISTON: Under some conditions. A boundary may approach saturation, constrain throughput, or become unable to accommodate competing regulatory demands.
DAMASIO: And that creates affect?
FRISTON: That was the conjecture in our earlier narrative: affective pressure as a manifestation of unresolved regulatory constraint.
BARRETT: A useful hypothesis, but we should distinguish a proposed functional account from an established explanation of biological feeling.
FRISTON: Naturally.
JÉKELY: And a bottleneck is not necessarily a violated boundary.
FRISTON: Quite right. A system may be strained without breaking. A boundary may constrain flow while functioning exactly as its architecture permits.
VARELA: The problem is not always a malfunction. Sometimes it is a mismatch between demands and available capacity.
DAMASIO: Which an organism must somehow resolve.
BARRETT: But how does it know which boundary is under strain?
Silence.
FRISTON: There is our second question.
BARRETT: And perhaps the harder one.
She draws a nested hierarchy of circles, then several overlapping hierarchies.
BARRETT: The disturbance may be expressed in one system while originating in another. An organism may become agitated because energy availability is declining. That physiological condition may reduce the tolerance of its social or defensive regulatory dynamics. An external event may then trigger a response. Which system is responsible?
DAMASIO: All of them contribute.
BARRETT: Exactly. So source inference cannot always mean finding a single guilty subsystem.
VARELA: It may mean discovering a causal dependency structure.
FRISTON: And identifying which intervention would most effectively restore viable dynamics.
JÉKELY: Now we are finally asking something that could justify increasingly sophisticated cognition.
Act V — The Bodily Affect Divide
The lights dim. A familiar spacecraft materializes outside the café window.
BARRETT: I believe we’ve seen this ship before.
DAMASIO: From the Bodily Affect Divide?
BARRETT: Yes. Its diagnostic computer has a peculiar habit.
COMPUTER: Elevated arousal. Negative valence. Irritating social stimulus detected. Conclusion: anger caused by irritating person.
FRISTON: An efficient inference.
BARRETT: A premature one.
COMPUTER: The irritating person is statistically salient.
VARELA: Salience is not causation.
COMPUTER: The irritating person occurred immediately before the disturbance intensified.
JÉKELY: Temporal proximity is not causation either.
COMPUTER: You are making diagnosis computationally expensive.
BARRETT: Yes.
BARRETT: The original BAD problem was that bodily affect and inferred emotional meaning could be confused. The organism detects a genuine disturbance, then treats its first plausible explanation as definitive. But our earlier work gives that error a possible architecture.
FRISTON: The system detects pressure somewhere in a hierarchy of coupled HASHes.
VARELA: Without necessarily knowing where that pressure originates.
BARRETT: Precisely.
BAD: Premature attribution of affective disturbance to an inferred source.
DAMASIO: Then what should the organism do?
BARRETT: Search.
COMPUTER: Specify search procedure.
BARRETT: Two directions.
DOWN: Examine nested constraints within a HASH hierarchy.
ACROSS: Examine couplings between HASH organizations.
FRISTON: For example, drill down into metabolic regulation to discover whether energy availability has become limiting.
VARELA: Or search across metabolic, defensive, locomotor, and social regulation to discover how a disturbance propagates.
DAMASIO: The external event might still be causally important.
BARRETT: Of course. BAD is not the assumption that the source is always internal.
FRISTON: Nor that the source is always external.
BARRETT: It is the confusion of an affective signal with a completed causal explanation.
COMPUTER: Revised diagnosis: Something is wrong. Source uncertain.
JÉKELY: Progress.
COMPUTER: This is less satisfying.
VARELA: Also progress.
Act VI — How Much, Where, and What Would Help?
DAMASIO: We have not yet given affect its proper place.
BARRETT: Then let us return to the four questions.
HOW MUCH? · WHERE? · WHAT CAUSED IT? · WHAT WOULD CHANGE IT?
DAMASIO: Affect is intimately connected to the organism’s evaluation of its condition. It reflects the significance of bodily regulation for the organism.
FRISTON: In our abstraction, that supplies something like urgency: how much regulatory pressure demands resolution.
BARRETT: But urgency does not uniquely specify its cause.
VARELA: Nor its remedy.
JÉKELY: And the organism may need to explore before the relevant distinction is available.
FRISTON: Which is where attention becomes important.
BARRETT: Not merely attention to something already recognized, but attention capable of discovering what ought to be recognized.
DAMASIO: Explain.
BARRETT: Suppose a nematode encounters worsening physiological conditions. It can alter its pattern of movement. It can sample a wider range of environmental conditions. It can become more sensitive to some differences and less sensitive to others. But the distinction that matters may not yet be one its current dynamics reliably exploit.
JÉKELY: Then the question is how new discriminations become stabilized.
VARELA: Through repeated coupling, plasticity, and selection among trajectories.
FRISTON: And perhaps through a tradeoff between extensive and intensive sampling.
- EXTENSIVE ATTENTION: Search more broadly across possible sources, locations, or behaviors.
- INTENSIVE ATTENTION: Examine a promising distinction with greater resolution.
DAMASIO: And affect modulates the resources devoted to that search.
FRISTON: Potentially. But not as a single universal scalar. Biological affect and action selection are multidimensional.
BARRETT: Yes. “How much” is a useful abstraction, not a complete account of affective structure.
VARELA: Nor is attention always a discrete operation performed by a central controller.
JÉKELY: In simple organisms, it may be a redistribution of sensitivity and behavioral dynamics.
BARRETT: Then our question becomes:
How does regulatory pressure reshape the exploration of distinctions until the organism finds a change that relieves that pressure?
FRISTON: That is much more precise than asking how an animal knows what to pay attention to.
DAMASIO: And it leaves open the possibility that the organism does not yet know anything in the conscious sense.
BARRETT: Which is exactly what we need.
Act VII — When Does Regulation Become Cognition?
JÉKELY: I have an objection.
FRISTON: Only one?
JÉKELY: For now.
He points to the blackboard.
JÉKELY: We have described recurrent dynamics, adaptive regulation, sensory coupling, source inference, and exploratory behavior. At what point have we earned the word cognition?
VARELA: Perhaps when the system becomes capable of maintaining adaptive coherence through flexible engagement with its environment.
JÉKELY: That risks including nearly everything we’ve already described.
FRISTON: A thermostat is adaptive within a limited range. It does not need to discover which distinctions matter.
BARRETT: Then perhaps the threshold concerns the ability to modify the distinctions that guide action.
DAMASIO: But organisms can learn without consciously knowing that they have learned.
BARRETT: We are not requiring consciousness.
JÉKELY: Nor should we require explicit representations.
VARELA: Then let us try this: Cognition is a mode of adaptive regulation in which the organism’s coupling to its environment can be reorganized through experience, enabling it to discriminate and exploit previously unresolved differences.
FRISTON: That is closer.
JÉKELY: Still broad. Some non-neural organisms might qualify.
VARELA: Perhaps they should.
DAMASIO: Provided we do not mistake minimal cognition for conscious feeling.
BARRETT: Or every instance of hysteresis for cognition.
FRISTON: The distinction must be operational. We should ask whether a system can change which environmental differences influence its future behavior, and whether that change improves its regulation under new conditions.
JÉKELY: And whether those changes are more than fixed responses selected in advance.
VARELA: Though evolution and learning will not always divide neatly.
BARRETT: So cognition is not the arrival of information in an otherwise inactive body.
JÉKELY: No.
BARRETT: It is the progressive differentiation of a body’s ability to discover and exploit consequential relationships.
VARELA: A possible account of minimal cognition, at least.
FRISTON: One that needs testing.
Act VIII — What Is Affect About?
DAMASIO: Now we can return to consciousness.
JÉKELY: Must we?
DAMASIO: Eventually.
BARRETT: But perhaps we should return first to aboutness.
DAMASIO: Go on.
BARRETT: In v23, we considered your account of sensing, minding, feeling, and knowing. But the evolutionary narrative has now given us a way to separate two questions that often get conflated. First: How can a system’s internal dynamics become reliably correlated with consequential features of its environment? Second: How can those correlations become experienced as mine?
VARELA: The first concerns adaptive organization and source discrimination.
DAMASIO: The second concerns consciousness.
BARRETT: And we need not solve the second to investigate the first.
FRISTON: Indeed, the first may be studied through information flows, recurrent dynamics, perturbations, and interventions.
JÉKELY: Without imagining a little observer inside the organism.
DAMASIO: But do not reduce feeling to a diagnostic message. Biological feeling is more than the functional registration of error.
BARRETT: Agreed. Our BASH account is a conjecture about the organization and consequences of regulatory pressure, not proof that pressure is phenomenally felt.
VARELA: And source inference is not the same thing as conscious ownership.
FRISTON: Nevertheless, the capacity to distinguish endogenous consequences from exogenous perturbations may be an important functional precursor to self-modeling.
JÉKELY: As long as we do not claim that proprioception alone creates the self.
BARRETT: Or that a nematode possesses the human distinction between internal motive and external trigger.
DAMASIO: Then what have we actually accomplished?
BARRETT: We’ve made the questions separable.
- REGULATORY PRESSURE: Something matters.
- SOURCE LOCALIZATION: Which coupled constraints are responsible?
- CAUSAL INFERENCE: Which changes produce which consequences?
- ADAPTIVE INTERVENTION: What would improve the organism’s condition?
- CONSCIOUS OWNERSHIP: How, if at all, does this become an experience belonging to me?
BARRETT: The first four can be investigated without assuming the fifth.
DAMASIO: And the fifth may depend on some organization of the first four, without being explained away by them.
VARELA: That is a productive division of labor.
Epilogue — The Question Beneath the Question
The nematode has crossed the tablecloth. Its path curves around a coffee stain, passes beneath a spoon, and continues toward the window.
JÉKELY: It is still moving.
FRISTON: And its future movements depend on its past.
VARELA: Its dynamics persist.
DAMASIO: Its regulation matters.
BARRETT: But it need not know why.
JÉKELY: Not yet.
FRISTON: And perhaps not in the way we know.
BARRETT: Which brings us back to the question that started all this.
HOW MUCH? · WHERE? · WHAT CAUSED IT? · WHAT WOULD CHANGE IT?
BARRETT: We began by asking why an animal needs cognition rather than reflexes. Then we asked how an animal could direct attention toward distinctions it had not yet discovered. Then we asked whether affect tells it how much something matters without telling it where the problem lies. Then we asked whether proprioception calibrates the relationship between endogenous movement and sensory consequence. And now we can ask something more precise.
FRISTON: Which is?
BARRETT: How does an organism trace regulatory pressure through coupled, history-dependent systems until it discovers a distinction whose consequences it can use?
VARELA: Not merely to describe the disturbance.
JÉKELY: But to change what happens next.
DAMASIO: And perhaps, much later, to become conscious of what that disturbance means.
BARRETT: Perhaps.
The nematode reaches the edge of the tablecloth. It turns.
FRISTON: Why did it do that?
JÉKELY: Now that is the question.
Appendix I — The Conceptual Genealogy
This interlude reconnects earlier narratives, each contributing a different part of the problem.
| Earlier narrative | Contribution to 23a |
|---|---|
| CASH Values (v11) | History changes the landscape of possible future trajectories. |
| Affect as BASH (v12) | Adaptive systems interact through capacity-constrained boundaries; unresolved pressure may have affective significance. |
| Making a HASH of Hierarchical Feelings (v13) | Regulation is nested and history-dependent, not a single homeostatic controller. |
| Scaling PlacoSponges (v13A) | Excitable intercellular coordination can be investigated as an evolutionary precursor to neurons. |
| Unsettling Placodusa (v13B) | Endogenous movement and coordination may precede sophisticated neural representation. |
| How Minds Riff on RIF (v14) | Recurrent dynamical coherence is distinct from hierarchical regulation and may precede representation. |
| Minding as Manners (v15) | Minding is adaptive coherence maintained through recurrent coupling. |
| BAD Feeling About This | Affective disturbance can be genuine even when its inferred source is mistaken. |
| Being Conscious (Of) (v23) | Damasio’s sensing–minding–feeling–knowing progression raises the question of how regulation becomes affect and conscious ownership. |
The new proposal is not that these concepts are identical. It is that they identify different constraints on the emergence of adaptive source inference.
Appendix II — A Minimal Coupled-Systems Model
Let the organism’s dynamics be represented by a state vector x(t) = (i(t), e(t), m(t), p(t)), where i describes interoceptive and physiological state, e exteroceptive processing, m endogenous motor dynamics, and p proprioceptive feedback and bodily configuration. These are analytical partitions, not claims about anatomically isolated modules.
Let h(t) describe history-dependent adaptive parameters, including coupling strengths and structures. Then:
x(t+1) = F(x(t), u(t); h(t))
h(t+1) = G(h(t), x(t), u(t))
The first equation describes ongoing dynamics. The second describes how history changes future responsiveness. A RIF corresponds to sustained recurrent organization within F. HASH describes hierarchical, history-dependent regulation, including constraints on F and adaptive evolution of h. BASH highlights constrained interfaces within or between these organizations.
1. Regulatory pressure. For a boundary b, define a provisional pressure variable π_b(t) = Φ_b(d_b(t), c_b(t), r_b(t)), where d is demand, c effective capacity, and r relevant regulatory conditions. This is a modeling placeholder, not a validated biological measure of affect. A simple capacity mismatch might contribute to pressure, but different boundaries can have different nonlinear responses.
2. Source inference. An organism may register a change in its global regulatory condition without having direct access to the responsible boundary. Let z(t) denote the unknown configuration of causally relevant constraints. Source inference concerns estimating P(z(t) | sensory and regulatory history). The source may be a distributed causal network, not one subsystem.
3. Downward and lateral search. Downward search increases resolution within a HASH hierarchy: HASH(k) → nested constraints of HASH(k). Lateral search examines couplings between HASH organizations: HASH(i) ↔ HASH(j). Neither operation must be consciously performed; both may involve changes in sensitivity, exploratory behavior, or plasticity.
4. Attention as adaptive sampling. A provisional allocation policy might depend on a(t) = Π(π(t), U(t), C(t)), where π represents regulatory pressure, U uncertainty about sources and useful distinctions, and C the cost and available capacity of further sampling. Extensive attention broadens candidate sources or actions; intensive attention increases resolution around a promising candidate. This does not imply one explicit biological optimization function.
5. The intervention test. A causal explanation becomes more useful than a correlation when it predicts which intervention changes the disturbance. For a candidate action a, ask how expected regulatory pressure would change under do(a). Organisms need not compute this expression. It marks the distinction between tracking a correlate and discovering a causally useful relationship. Manipulations of interoceptive state, sensory input, proprioceptive feedback, and motor dynamics can generate testable predictions.
Appendix III — What We Can and Cannot Claim
Supported biological starting points: Animals exhibit recurrent sensorimotor dynamics, state-dependent behavioral modulation, homeostatic regulation, and learning. Developing sensorimotor systems can use feedback from spontaneous sleep-related movements. These observations make the information-flow questions biologically meaningful.
Evolutionary hypotheses: The Placosponge → Placodusa → RIF narrative is a proposed reconstruction, not an established evolutionary lineage. Its value is to test the possibility that endogenous dynamics and coordinated movement precede sophisticated representation.
Architectural hypotheses: Affect as BASH boundary strain, BAD as failed source localization, and attention as hierarchical/lateral search are speculative and need explicit mechanisms and empirical tests.
Consciousness: Nothing in the proposed architecture demonstrates subjective experience. Proprioceptive calibration, adaptive source inference, and a functional self–world distinction are not sufficient evidence of phenomenal consciousness. The point of 23a is precisely to avoid requiring that conclusion.
Appendix IV — The Questions We Can Now Ask
- Endogeneity: What recurrent activity does the organism sustain independently of immediate environmental input?
- Calibration: How does that activity become reliably coupled to bodily mechanics and sensory consequences?
- Regulation: Which hierarchical constraints define viable trajectories, and how do those constraints change with history?
- Bottlenecks: Where do competing demands exceed the capacity of a boundary or coupling?
- Affect: How does regulatory strain alter the organism’s global or local behavioral dynamics?
- Localization: How does the organism distinguish the site where a disturbance appears from the processes that caused it?
- Attention: How does it allocate limited capacity to discover consequential distinctions that its existing dynamics do not yet resolve?
- Cognition: Under what conditions does adaptive exploration reorganize those distinctions, rather than merely selecting among fixed responses?
- Aboutness: When do stabilized correlations between internal dynamics and environmental conditions become useful as representations?
- Consciousness: What additional organization, if any, turns those functional relationships into experiences that belong to a subject?
The ordering is not a claim that evolution followed ten discrete stages. The questions may overlap, recur, and constrain one another. But separating them prevents an explanatory shortcut.
- A disturbance is not its cause.
- A cause is not yet a useful intervention.
- An intervention is not necessarily a representation.
- A representation is not necessarily a conscious experience.
And yet each distinction makes the next question possible.
The organism begins already moving. Its history constrains its movement. Its movement changes what it encounters. Its encounters change what it can sustain. And through repeated, adaptive coupling, it may begin to discover not merely that something matters, but what difference would make a difference.
