The Hypervisor Switching Law¶

The Problem¶

Each of the 12 dimensions is a semi-autonomous agent (a 'mind')
They exist in a MULTISET (a bag) — not ordered, not ranked a priori
Each one has its OWN WILL — all equally valid
ANY of them could be the hypervisor at ANY time
They are CONSTANTLY trying to replace each other
There must be a MATHEMATICAL LAW that determines which one gets priority
This law is the core of the advanced game theory (Sigma-21)

The Mechanism: Contextual Fitness Auction¶

Formalization¶

Let $d_i$ for $i = 1, \ldots, 12$ be the twelve dimensions.

At each moment $t$, the environment presents a stimulus vector: $$\mathbf{s}(t) \in \mathbb{R}^m$$

Each dimension $d_i$ computes a FITNESS SCORE — how relevant it is to the current stimulus: $$f_i(t) = \langle \mathbf{w}_i, \mathbf{s}(t) \rangle \cdot \alpha_i(t) \cdot v_i$$

Where: - $\mathbf{w}_i$ = dimension $i$'s weight vector (what stimuli it responds to) - $\alpha_i(t)$ = dimension $i$'s current activation level (how developed it is) - $v_i$ = dimension $i$'s individual will (its drive to become hypervisor)

The Switching Rule (Softmax Selection)¶

The probability of dimension $d_i$ becoming hypervisor at time $t$:

\[P(H_t = d_i) = \frac{e^{\beta f_i(t)}}{\sum_{j=1}^{12} e^{\beta f_j(t)}}\]

Where $\beta$ is the INVERSE TEMPERATURE: - High $\beta$ (cold): winner-take-all, sharpest switching, one dimension dominates - Low $\beta$ (hot): distributed control, multiple dimensions share influence - $\beta = 0$: uniform distribution — all dimensions equally likely (chaos) - $\beta \to \infty$: deterministic — highest fitness always wins (rigidity)

The Healthy Range¶

A healthy intelligence operates at INTERMEDIATE $\beta$
Sharp enough to select the right hypervisor for the context
Soft enough to allow smooth transitions and multi-dimensional awareness
TRAUMA can push $\beta$ too high (one dimension stuck in control — hypervigilance)
DISSOCIATION can push $\beta$ too low (no dimension takes control — chaos)

Five Candidate Switching Laws¶

1. Softmax (Boltzmann) — described above¶

Probabilistic, smooth transitions
Used in neural network attention mechanisms
Natural temperature parameter
Most likely candidate for biological systems

2. Winner-Take-All (Argmax)¶

$$H_t = \arg\max_i f_i(t)$$ - Deterministic, sharp switching - No blending, no gradual transitions - Too rigid for biological reality — but may approximate high-stress states

3. Proportional Control (Linear)¶

$$P(H_t = d_i) = \frac{f_i(t)}{\sum_j f_j(t)}$$ - Simpler than softmax, no temperature parameter - More sensitive to small differences - May be the 'resting state' when no strong stimulus is present

4. Threshold + Tournament¶

Only dimensions with $f_i(t) > \theta$ (threshold) enter the competition
Among qualifying dimensions, softmax or argmax selects the winner
The threshold $\theta$ is the 'attention filter'
Dimensions below threshold cannot bid for hypervisor
May explain why collapsed dimensions (trauma) can't compete

5. Hysteresis (Sticky Switching)¶

$$H_t = \begin{cases} H_{t-1} & \text{if } f_{H_{t-1}}(t) > \max_{i \neq H_{t-1}} f_i(t) - \delta \\ \arg\max_i f_i(t) & \text{otherwise} \end{cases}$$ - Current hypervisor keeps control unless another dimension exceeds it by margin $\delta$ - Prevents rapid oscillation (chattering) - $\delta$ = switching cost / inertia - Explains why people get 'stuck' in one mode even when context changes

The Real Law: Probably a Combination¶

The biological switching law is likely: 1. Threshold filter (only activated dimensions can compete) 2. Softmax selection (among competitors, probabilistic based on fitness) 3. Hysteresis (current hypervisor has inertia, resists switching) 4. Will override (the meta-Will can force a switch regardless of fitness scores)

Combined: $$H_t = \begin{cases} W(t) & \text{if Will override active} \\ H_{t-1} & \text{if } f_{H_{t-1}}(t) > \max_{i: \alpha_i > \theta} f_i(t) - \delta \\ \text{Softmax}(\{f_i(t) : \alpha_i(t) > \theta\}) & \text{otherwise} \end{cases}$$

Connection to Existing Framework¶

Trauma¶

Trauma collapses a dimension's $\alpha_i$ below threshold $\theta$
That dimension can no longer compete for hypervisor
The remaining dimensions have less competition → narrower switching → more rigid behavior
Recovery = raising $\alpha_i$ back above threshold → dimension re-enters the auction

Maximum Activation¶

All 12 dimensions above threshold = maximum competition = richest switching
The hypervisor changes fluidly based on context
This is cognitive FLEXIBILITY — the ability to deploy the right mind for the right moment
12/12 activation = the full multiset participating in the auction

The Will¶

Will is the META-HYPERVISOR — it can override the automatic auction
'I am going to focus on this' = Will forcing a specific dimension to hold control
Will override is expensive (costs energy) but powerful (directed motion)
Without Will override: the system runs on autopilot (Brownian motion)
With Will override: the system follows a directed path (world line)

Dyscalculia Explained¶

Mathematical (computational) dimension: low $\alpha$, below threshold for symbolic tasks
Abstract (structural) dimension: high $\alpha$, wins hypervisor for pattern recognition
When faced with algebra: computational dimension can't compete → Abstract takes over
Abstract sees structure but can't compute → the 'can do postdoc math, can't do undergrad math' pattern

Open Questions for Sigma-21 Math¶

What determines $\beta$ (the temperature)? Neurochemistry? Arousal? Development?
Is $\beta$ the same for all dimensions or dimension-specific?
Can the threshold $\theta$ be measured empirically?
What is the switching time constant? (How fast can hypervisor change?)
Is there a topological constraint on switching? (Can any dimension switch to any other, or are some transitions forbidden?)
How does the neurochemical stack affect $\beta$ and $\theta$?
Can we observe hypervisor switching in fMRI data?