Most systems that evaluate anything — health interventions, product quality, workplace performance, life satisfaction — use some form of weighted average. Take each factor, assign it a weight, multiply, sum. It's intuitive, simple, and often insufficient.
It's insufficient because it assumes the factors are independent. That improving sleep has no effect on how much benefit you get from exercise. That nutritional adequacy doesn't change the impact of cognitive training. That attachment security is unrelated to stress regulation capacity.
The evidence says otherwise. But the strength and precision of what we can claim deserves honesty.
What's well-evidenced: specific synergies
Certain pairwise interactions between dimensions are documented through identified biological or psychological mechanisms:
| Interaction | Mechanistic basis |
|---|---|
| Attachment security × Stress regulation Relationship Quality |
Secure attachment provides neurobiological infrastructure for stress regulation — co-regulation with a trusted partner reduces amygdala activation (Coan's fMRI studies). Without attachment security, stress regulation relies entirely on self-regulation, which is more metabolically costly and less effective. |
| Sleep duration × Circadian alignment Sleep Optimization |
Leproult 2014: circadian misalignment DOUBLED metabolic disruption beyond sleep restriction alone. Eight hours at the wrong circadian phase produces measurably worse outcomes than eight hours at the right phase. These are two independent biological axes that multiply harm when both are degraded. |
| Attention regulation × Emotional regulation Contemplative Practice |
You cannot regulate what you cannot attend to. Attention regulation is the prerequisite for emotional regulation. Contemplative traditions have known this for millennia — concentration (samādhi) precedes insight (paññā). |
| Aerobic base × Neurobiological response Physical Activity |
The aerobic base IS the stimulus; the neurobiological response IS the adaptation. Erickson 2011 directly demonstrated the causal chain: aerobic training → BDNF increase → hippocampal volume increase → spatial memory improvement. |
These interactions are individually well-supported. The mechanism is identified, the evidence is cited, and the direction (synergistic — both must be strong for full effect) is documented.
What we overclaimed: the 4:1 ratio
Our earlier analysis reported a clean pattern: approximately 4 synergies for every 1 redundancy across all branches. We presented a table showing this ratio as consistent across every domain and described it as "a structural property of how human needs relate to each other."
When we audited the underlying data, we found something important: every single framework has exactly 1 redundancy. The ratios are 3:1, 4:1, or 5:1 — but always with precisely one redundant interaction. This uniformity is almost certainly an artifact of how the frameworks were generated (an LLM-based framework deriver that likely had a structural prior for including one redundancy per framework), not a discovered property of human biology.
The overall ratio (approximately 3.79:1 synergistic to redundant) is real in the data we have. But the data was shaped by the generation process. We can say:
- The specific synergistic interactions listed above are individually well-evidenced
- The biological mechanisms for non-independence are documented
- The overall pattern skews synergistic (more amplification than substitution)
We cannot say the precise ratio is a discovered law of human needs. That requires interaction indices derived from empirical data, not from LLM-generated frameworks — work that hasn't been done yet.
What synergy means for scoring — honestly
In our computed scoring model, we applied conservative interaction adjustments of ±0.05 based on direction only (synergy vs. redundancy), not on the LLM-estimated magnitudes. The effect: synergies shift scores by approximately 5-10%.
This is meaningful but not transformative. It's a refinement, not a revolution. The earlier article's claim that "a scoring system that ignores this misses 80% of the interaction structure" was based on counting interactions, not on measuring their quantitative impact on scores.
The honest mathematical difference: in a weighted average, if sleep scores 0.8 and inflammation scores 0.8, the combined contribution might be 0.48. With our conservative interaction adjustment, it becomes approximately 0.50-0.52. Not the dramatic "0.76" we previously calculated — that figure used interaction indices we now know were too strong for valid Choquet computation (all 19 frameworks had floor sums ≥ 1.0, making 2-additive Choquet technically invalid).
Human needs are non-independent. The evidence for specific synergistic interactions is individually strong and mechanistically grounded. The overall pattern skews synergistic. But the precise ratio, the magnitude of interaction effects on scores, and the uniformity across domains are less certain than we originally presented. The interactions matter — but they're a second-order effect on top of the first-order importance of each dimension's research-grounded effect size.
What does hold up
The practical implication survives: don't optimize dimensions in isolation. The evidence for specific synergies — sleep and circadian alignment, attachment and stress regulation, attention and emotional regulation — is individually robust. Getting three related dimensions to "good" likely produces more flourishing than getting one to "excellent" because the documented interactions between them are real.
But we should hold this as "the evidence supports non-independence" rather than "we've precisely quantified the ratio." The former is well-grounded. The latter requires empirical interaction measurement we haven't yet performed.
What redundancy tells us
The redundant interactions, while potentially over-counted in our framework generation, point to a real efficiency principle where they're individually documented:
- Whole-diet pattern × Anti-inflammatory nutrition: Mediterranean and MIND diets already include the major anti-inflammatory components. Targeted supplementation adds less value when whole-diet quality is already high.
- Breathwork × HRV biofeedback: Both target the same autonomic pathway — vagal tone enhancement through controlled respiratory patterns. The physiological mechanism overlaps.
- ACT therapy × Digital self-guided programs: Both deliver the same core intervention. The therapist-delivered version is more effective but the content substantially overlaps.
These tell you where investing in both produces diminishing returns — useful for efficiency, even if the precise magnitude is uncertain.