CLM-L001 — Alignment Index (AX)
Status: ✅ Integrated · 🔍 Practitioner-grounded · Falsifiable ✓ — locked into THEORY-OF-TRAITS.md 2026-04-29 (state-vs-capacity layer beneath AQ)
Structuring status: Pre-adoption Locked / In structuring
Confidence: Locked (under historical informal protocol; pending GOV-24 re-evidencing)
Last reviewed: 2026-04-29
Topic: 01-alignment-metrics
CLAIM TEXT
The Alignment Index (AX) is a state-based metric (0–100) representing how aligned a person is right now. It is computed as the simple average of two sub-scores: TAS (Trait Alignment Score, how fully core traits are engaged) and SAS (Situational Alignment Score, how supportive the environment is). AX answers a different question from AQ: AX = "where are you now?", AQ = "how skillfully can you manage being there?"
LOCATION (pre-adoption)
archive/planning-desk/RAG & Articles/Markdown Articles/App and Dev Files/Building Birthday Bot/Alignment Index (AX).md (v1.0, 2025-04-30, Steven Rudolph)- Cross-referenced in
Master List of Parameters for Alignment Tracking System.md, 06.3_Alignment_Parameter_Spec.md (Brain Upgrade 2025-05-05)
LOCATION (post-adoption, when integrated)
Not yet integrated into knowledge/theory/THEORY-OF-TRAITS.md. Current canon names AQ but does not name AX/TAS/SAS as the underlying state-metric layer.
EVIDENCE TYPES
[P] Phenomenological
Practitioner observation across many cases that "where you are right now" and "how well you can manage it" are different layers. Clients can have intact reading-and-decision capacity (AQ) while still being in deep current misalignment (low AX) — and vice versa: clients can be situationally well-placed (high AX in the moment) without the capacity to sustain or restore it under change (low AQ). The AX/AQ split mirrors the "clear-eyed paralysis" pattern locked in the AQ canon (high AQ × low will). N: hundreds of cases over 18+ years; specific N for AX-vs-AQ separation observation: not formally tracked.
[E] Empirical
- rs-0014 (Kahneman, 2011) — supports the dual-process basis on which AX (state) and AQ (capacity) sit. Light backing only — Kahneman's System 1/2 split is about cognitive process, not alignment-state vs. alignment-capacity. The analogy is suggestive, not direct.
- MISSING — direct empirical literature on state-vs-capacity distinction in psychometric measures. Possible candidates: trait-vs-state anxiety (Spielberger STAI); momentary affect vs. dispositional affect; trait emotional intelligence vs. ability EI. Need rs- entries.
- MISSING — psychometric validation literature for composite-average scoring (TAS + SAS / 2). Simple averaging is a strong assumption; real-world validation needed.
[T] Theoretical
Coherent with current theory's three-axis model (Nature × Situation × Orientation):
- TAS = the Nature-on-current-Situation read (which traits are activated, how well)
- SAS = the Situation-on-current-Nature read (whether the environment supports trait expression)
- AX = the resultant state at this intersection
- AQ = the developed quality of the Orientation axis (how skillfully the person reads, chooses, and adjusts at the Nature × Situation intersection)
The 4-quadrant mapping in the original AX rubric (Peak Performance / Misfit Success / Struggle Zone / Potential Blocked) follows directly from the TAS×SAS grid.
[C] Convergent
- MISSING — Self-Determination Theory (Deci & Ryan) names autonomy/competence/relatedness as situational supports for intrinsic motivation. SAS is functionally similar to SDT's "autonomy-supportive environment." Worth an rs- entry.
- MISSING — Csikszentmihalyi flow research treats flow as a state at the intersection of skill and challenge. AX is structurally analogous (state at the intersection of trait-readiness and situation-support). Worth an rs- entry.
- MISSING — Person-Environment fit literature (industrial/organizational psychology, e.g. Kristof-Brown, Edwards). Direct convergent-evidence candidate.
UPSTREAM SOURCES
- Steven Rudolph (2025-04-30). Alignment Index (AX) Scoring Rubric, v1.0. Internal Xavigate document.
- Self-source: practitioner-derived; not yet externally peer-reviewed
- Kahneman, D. (2011). Thinking, Fast and Slow. Farrar, Straus and Giroux.
- rs-0014 (assumed; verify)
- Used analogically; not direct support
- Schneider, W., & Shiffrin, R. M. (1977). Controlled and automatic human information processing. Psychological Review, 84(1), 1–66.
- rs- ID: verify; cited in current canon for native vs. forced activation, applies indirectly here
POSITIONING IN LITERATURE
- Confirms: The state-vs-trait distinction in psychological measurement (long-standing in psychometrics). Person-environment fit traditions in I/O psychology. Ecological validity arguments in psychology.
- Extends: Names the specific TAS×SAS structure as the operationalization of "alignment as a state." Most frameworks treat alignment as a vague gestalt; AX makes it computable.
- Departs: From frameworks that conflate state and capacity (e.g., simple "well-being" indices that bundle current state with stable disposition). AX explicitly separates them, with AQ as the capacity layer.
FALSIFIABILITY
The claim that AX = (TAS + SAS) / 2 produces a meaningful summary score would be falsified if:
- Composite AX scores predicted outcomes worse than either TAS or SAS alone in field use (the average destroys information rather than preserving it)
- Real-world cases regularly showed AX moving independently of both TAS and SAS (signaling missing variables)
- The 4-quadrant mapping failed to discriminate between meaningfully different practitioner cases
The state-vs-capacity distinction (AX ≠ AQ) would be falsified if: cases where AX and AQ moved in lockstep across all observed conditions (suggesting they measure the same underlying construct).
EDGE CASES / KNOWN LIMITS
- Quadrant IV ambiguity — Original rubric notes Quadrant IV (high TAS, low SAS = "Potential Blocked") may score numerically like Quadrant II (high SAS, low TAS = "Misfit Success") due to averaging. Original rubric calls for "secondary pattern detection." This is a known weakness — the average loses the asymmetry.
- Self-report dependence — TAS and SAS scoring relies on user self-report or AI-inferred language. No external behavioral validation built in.
- Cross-cultural calibration — All practitioner cases backing this rubric are US/India/France. Norms may differ in other contexts.
- Threshold artifacts — Score-range cutoffs (76–100, 51–75, etc.) are descriptive, not validated. The boundaries are practitioner judgment, not data-derived.
DISCONFIRMING CASES TRACKED
None formally tracked under this rubric (legacy material). To structure: pull cases where AX prediction differed from observed practitioner reading, and characterize the pattern.
REFLEXIVITY NOTE
This metric was developed by Steven (the framework's originator) for use in the Xavigate Birthday Bot system in 2025-04. The averaging structure (TAS + SAS / 2) reflects an engineering choice — make it computable for an AI session — more than a theoretical claim that the two contribute equally. A weighted version (e.g., trait-dominant weighting under stable conditions, situation-dominant weighting under acute change) is more defensible theoretically but was not implemented. The legacy rubric's strength is operational; its weakness is psychometric.
RELATIONSHIP TO CURRENT CANON
- Already integrated? No
- Contradicts current canon? No — but the AX/TAS/SAS structure is net-new to current canon. THEORY-OF-TRAITS.md names AQ but does not name AX as the state layer beneath.
- Net-new? Yes (in current canon's vocabulary, though present in legacy material)
- Recommended action: Cherry-pick into current canon. The state-vs-capacity layer matters for the framework's coherence:
- AX = state (where you are now)
- AQ = capacity (how well you manage being there)
- The current AQ section in THEORY-OF-TRAITS.md gains precision when AX is named beneath it
- The 4-quadrant mapping is useful for practitioner training (maps cleanly onto "clear-eyed paralysis," "forced action," "Wu Wei," "drift" already locked in current canon under the AQ × will table)
Suggested integration point: add a subsection to the AQ canon — "AX (state) and AQ (capacity) — two layers, not one metric." Reference legacy formulation, note its operational origins.
RESEARCH-BANK GAPS FLAGGED
For RESEARCH-BACKLOG.md:
- State-vs-trait psychometrics literature — Spielberger STAI (state-trait anxiety inventory) as a model; Larsen & Kasimatis (1990) on dispositional vs. momentary affect; rs- entry needed.
- Person-Environment fit literature — Kristof-Brown et al. (2005) meta-analysis on PE fit; Edwards & Shipp (2007); rs- entry needed.
- Self-Determination Theory — Deci & Ryan (2000) and follow-ups on autonomy-supportive environments; rs- entry needed for SAS theoretical grounding.
- Csikszentmihalyi flow — already convergent in current canon, but not formally cited as rs-entry; verify and add.
- Composite scoring validation — psychometric literature on simple-average composites vs. weighted composites; needed to defend the (TAS+SAS)/2 structure.
NOTES
- The original rubric refers to "ASS — Alignment Stability Score" as a downstream longitudinal metric (AX trends over time). ASS does not have its own dedicated source file in the audit; treated as part of the AX-trends discussion. Worth a separate claim file (CLM-L005 or similar) once source is confirmed.
- The metric was designed for the Xavigate AI session system (Birthday Bot, 2025). Its operational context shapes its form: a simple, computable score for real-time AI use. Theoretical refinement (e.g., weighted TAS/SAS, Quadrant-IV asymmetry handling) was deferred to future work.
- Term "alignment" in AX/TAS/SAS is the same alignment locked in current canon (the canon technical term for trait-to-situation match). Continuity is clean.
- Lineage: The metrics emerged from legacy alignment work (2023–2024), which itself emerged from Quantum Dynamics (early 2023). The Quantum Dynamics to AX/AQ lineage is documented across multiple legacy files in
archive/Xavigate/LMS/LMS 0/.