A Refinement of Known Weakness #3 in Stake-Geometry
Developed in conversation, May 22, 2026 — extends stake_geometry.md
What This File Is
The stake-geometry document lists, as Known Weakness #3, the worry that its five dimensions may be ad hoc — chosen because they seem to carve the phenomenon at its joints, with no principled derivation showing why these five and not three or eight. This file refines that weakness in light of cross-framework work between stake-geometry and the five-axis emotion model. The original worry was vague. What replaces it is sharper, more tractable, and more honest.
The Original Weakness, Restated
Five dimensions — dimensionality, depth, reach, coherence, responsiveness — were selected because they appear to track structurally distinct features of how a being is embedded in reality. The selection is defensible on inspection. It is not derived. A determined critic can ask: why five? Why not collapse depth and responsiveness, or split reach into temporal and spatial components? The framework has no answer beyond "these seem to carve at the joints."
This is a real weakness, and it deserves a real response.
Stake-geometry's own statement of the weakness names two ways out: a derivation showing why these five are the natural axes of variation, or strong empirical evidence that they are. It is worth saying plainly which of the two this refinement supplies. It does not derive the dimensions, and it does not pretend to — the deeper completeness question (why this set rather than another) is left open, and named as such in the closing section. What it does is work the empirical avenue: it reframes the worry so that it becomes a well-posed, testable question, and it identifies the criterion by which that question would be settled. The contribution is a down payment on the second way out, not the first.
What Cross-Framework Work Made Visible
Applying the five-axis emotion model to seven emotions, and then mapping each emotion axis to one or more stake-geometry dimensions, surfaced a pattern the original weakness did not anticipate.
Some dimensions appear to track genuine structural joints. Depth and responsiveness behave like distinct capacities that cannot be reduced to each other. A being can be high-depth and low-responsiveness (a dogmatic person who feels intensely but cannot update), or high-responsiveness and low-depth (a flexible thinker who never quite cares deeply). These are not different amounts of one thing; they are different things.
Other dimensions appear to bundle sub-capacities that can vary independently. Reach is the clearest case. As currently defined, reach covers temporal reach (caring across time), spatial reach (caring across distance), and abstract reach (caring across levels of generality). These three can come apart in real beings. A person devoted to their grandchildren but indifferent to strangers has high temporal reach and low spatial reach. A mathematician who cares about abstract truths but is parochial in time and space has high abstract reach and little of the other two. The same name covers three capacities that don't always travel together.
The five-axis emotion model has the mirror structure. Four of its axes appear to track distinct features of an episode — intensity, granularity, processing level, components — and the fifth, object, bundles three sub-charts (temporal arrow, definiteness/agency stack, credence × stake heat-plot) whose behaviors can dissociate. Disgust depletes its temporal sub-block but keeps the credence × stake plot fully alive. Surprise depletes the heat-plot but the temporal sub-block, while collapsed to an instant, is still doing different work than definiteness. The three sub-charts can rise and fall independently, which is the empirical signature of bundled capacities rather than facets of one thing.
The Joint / Bundle Distinction
The refined version of the weakness is this. The dimensions identify real structural features of stake-geometry, but the level of granularity at which they are individuated is partly conventional. Two kinds of dimension are mixed together:
-
Joints — dimensions that carve structural reality at a genuine seam. Their components do not dissociate; varying one component varies them all together because they are facets of one underlying capacity. Likely joints in the current five: depth, responsiveness.
-
Bundles — dimensions that group structurally similar sub-capacities under one name for reasons of common function and pedagogical compactness. Their components can dissociate in real beings. Likely bundles in the current five: reach (temporal + spatial + abstract — where temporal reach means span, the distance stakes extend across time, not the decay-rate "temporal dynamics" the mood work introduced as a separate sub-dimensional layer), possibly dimensionality (which might bundle the variety of distinct objects with the depth of relational integration with other beings).
Coherence is the hardest to classify without further work. It plausibly bundles intra-episode coherence (components aligned within a single emotional event) with cross-episode coherence (values, projects, and relationships pulling in compatible directions over time). Whether these are facets of one capacity or empirically dissociable is an open question.
The mistake the original weakness statement implied was that the choice of five was either right or wrong. The truer picture is that the choice of five is partly right and partly conventional. Some of the five are discoveries about structure. Others are useful groupings whose members happen to share enough structural function to be named together.
Why This Refinement Is a Step Forward
The original weakness was hard to act on. "Maybe the five are right, maybe they aren't, no way to tell" admits no progress. The refined weakness is tractable in several ways.
First, it converts a conceptual worry into an empirical question. Whether reach's three sub-capacities correlate tightly across individuals — whether high temporal reach predicts high spatial and abstract reach — is, in principle, measurable. Same for the other suspected bundles. The framework now has a research program rather than a permanent caveat.
The evidence actually in hand, though, is of one kind only: illustrative dissociation. Cases like the dogmatic feeler (high depth, low responsiveness) or the parochial mathematician (high abstract reach, low temporal and spatial reach) show that two capacities can come apart. That establishes possibility, not frequency. The stronger claim the refinement needs — that a suspected bundle's components routinely dissociate across a population — requires measured correlational data the framework does not yet have. So the criterion is in place and the preliminary signs point the right way, but the data is still owed. This is the avenue opened, not the avenue walked.
Second, it preserves what is strongest in the current framework. The dimensions identified are real. Depth, responsiveness, dimensionality, reach, and coherence each track something structural about how a being is embedded in mattering. The refinement is not that the framework is wrong but that some of its categories operate at a coarser grain than others.
Third, it offers a principled response to the skeptic. The defense of the current count is no longer "it seems to carve at the joints." It is: "Some of the five appear to be joints — we name them and offer the evidence. Others are pragmatic bundles whose components share structural function — we name them as bundles and identify the sub-capacities they group. A future version of the theory may split the bundles if the empirical evidence shows their components routinely dissociate."
That is a defensible position rather than a deflection.
The Symmetry With the Emotion Model
The same joint-versus-bundle structure appears in the five-axis emotion model, and the parallel is not coincidence. Both frameworks were built to carve their respective phenomena at structural joints. Both succeeded at identifying real features. Both made some choices about granularity that were partly principled and partly pragmatic.
Stake-geometry's reach is the emotion model's object panel: a bundle of structurally similar but empirically separable sub-capacities, grouped under one name because the grouping is functionally and pedagogically clean. Stake-geometry's depth is the emotion model's intensity axis: a genuine joint, irreducible to anything else in the framework.
The symmetry suggests something important. The two frameworks are not independent attempts to map different phenomena. They are two scales of measurement on the same underlying structure — the lifetime geometry and the episode within it. That the same joint/bundle pattern shows up at both scales is consistent with their being two views of one thing.
This is also why the choice to bundle or split should likely be made consistently across both frameworks. If reach is split into temporal, spatial, and abstract, the symmetry argument says the object panel should be split into its three sub-charts as well. If reach stays bundled, object stays bundled. A mature version of the project will probably commit to one of these two configurations explicitly, with the reasons stated.
One asymmetry in the parallel is worth flagging, because it tells you where the evidence currently sits. The emotion-model side rests on closer-to-worked analysis: the claim that the object panel's sub-charts dissociate comes from actually applying the model across seven emotions and watching disgust deplete its temporal sub-block while surprise depletes the credence × stake heat-plot. The stake-geometry side, by contrast, still rests on illustrative cases. The episode scale has been exercised against worked material in a way the lifetime scale has not. If the symmetry argument holds, that asymmetry is an opportunity rather than an embarrassment: evidence accumulated at the episode scale is partial evidence about the lifetime dimensions those episodes aggregate into.
Implications for the Framework
A few practical consequences follow.
The stake-geometry document's Known Weakness #3 section should be rewritten to reflect this refinement. The original phrasing — "the dimensions may be ad hoc" — is replaceable with the more accurate "the dimensions identify real structural features but are individuated at a partly conventional level of granularity, with some functioning as joints and others as bundles."
Future empirical work on the framework should distinguish two questions. One is whether the five dimensions track real structure (the framework's central claim). The other is whether the level of granularity is correct (a separable question about taxonomic choice). Conflating these has made the ad-hoc worry feel more threatening than it is.
The bundles, once named, can be examined. Reach's three sub-capacities can be tested for empirical correlation. Coherence's internal structure can be examined across episodes and across lifetimes. Dimensionality can be checked for whether it really is one capacity or two. The work of sorting joints from bundles is the work of bringing the framework from defensible-on-inspection to defensible-on-evidence.
The project's own logged episodes are the natural first place to look. If the object-panel sub-charts can be read off the records already in 03_data_workbook/emotion_episodes.xlsx, a first correlational pass — do the sub-capacities dissociate, and how tightly — becomes possible without collecting new data, converting the promissory note above into at least a preliminary result.
What Remains Unresolved
This refinement does not solve the ad-hoc problem entirely. Three things remain open.
First, the assignment of specific dimensions to joint or bundle is provisional. Depth and responsiveness look like joints. Reach looks like a bundle. But without correlational data across many individuals, these are educated guesses, not conclusions.
Second, the joint/bundle distinction is itself partly continuous. A capacity might be a bundle whose components correlate at 0.6, which is enough to make the grouping meaningful but not so much that it constitutes a single thing. The framework will need to develop a finer-grained vocabulary for this.
Third, the deeper question — why this set of joints and bundles rather than some other — remains. Even if depth, responsiveness, dimensionality, reach, and coherence are correctly identified at the right level of granularity, there is no derivation showing that these are the complete set of structural features of mattering-systems. Candidates not currently in the framework (reflexivity, agency, receptivity, relationality) remain on the table. One candidate from an earlier version of this list — temporality as a standalone dimension — has since been examined and set aside: the mood work split "temporality" into four senses and found that temporal reach already lives inside the reach bundle while temporal dynamics belongs to a process-layer beneath the dimensions (the mood mesoscale), so neither warrants a sixth dimension (see Mood and Sentiment and The Bridge Between the Frameworks). The refinement narrows the ad-hoc problem; it does not eliminate it.
A later development bears on the dimensionality suspicion specifically. The mass model (core_mass_model.md) reads dimensionality and depth as two coordinates of a single object, the distribution of mattering: one is how many places the mass sits, the other is how much sits in each place. If that reading holds, dimensionality is not a standalone capacity that might bundle, but one face of a joint it shares with depth, and the bundle question for it partly dissolves rather than resolves. The model makes this a commitment to be tested, not a finding.
This document records a theoretical refinement developed through dialogue. It is exploratory and is offered as a contribution to thinking about the structural foundations of the stake-geometry framework.