M43 F05 The Descent Clarified: Floor Jets, the Shelf's Second Order, and the Ceiling

The Theory of Operations moves between ranks in two fundamentally different ways. The HyperCore’s elevator is a global construction: from a rung it builds the next channel, pair protocol, storage mean and operation. A floor or ceiling jet is instead a local limit of one mean near a boundary point. It records only the first few coefficients of a germ and may send distinct means to the same limiting object.

F05 "The Descent Clarified" separates these two mechanisms and applies boundary jets to the superlogarithmic shelf presentation at rank 7/2. It proves that the presentation’s floor is blind at first order, calculates the second-order term that distinguishes its legs, and derives a new ceiling law in which the presentation contracts to Gauss’s arithmetic-geometric mean. The resulting Elevator Shadow holds at the level of first-order jets but fails for the full germs in the identity chart.

Two ways of moving between ranks

Let C_R be a HyperCore operation. The elevator constructs rank R+1 through five global stages:

1. iterate C_R to form a channel;
2. take the channel’s diagonal;
3. use the inverse diagonal as a normaliser;
4. iterate the two normalised directed readings as a pair protocol;
5. take their common limit as the storage mean M_(R+1) and reconstruct the new operation from its diagonal.

This procedure builds one operation from another over an entire domain. It is not a boundary approximation and need not be invertible.

A floor jet is different. If a mean M satisfies

M(1+epsilon*a,1+epsilon*b)

= 1+epsilon*Fl_1[M](a,b)+epsilon^2*Fl_2[M](a,b)+...,

then Fl_k[M] is its k-th floor jet. At an upper endpoint L, the analogous expansion

M(L-epsilon*a,L-epsilon*b)

defines the ceiling jets Ce_k[M].

A jet describes one mean near one corner. It forgets most of the global construction and is generally many-to-one. Thus a mean of one rank may have a lower-rank mean as its first-order floor limit without providing a second construction between those ranks.

For the HyperCore’s own storage means, the first-order floor ladder is

AM, AM, GM, HM

at integer ranks 1,2,3,4, with the rank-four law still dependent on the stated floor hypothesis. At the half ranks, the Symmetric Hyperelliptic Mean SHM descends to Gauss’s arithmetic-geometric mean AGM, while the rank-7/2 HyperCore protocol ISHE-H conditionally descends to the reflected geometric-harmonic mean

GH(a,b) = a*b/AGM(a,b).

The elevator and the floor jet therefore encode different relationships. The elevator generates the hierarchy; the floor jet reveals how one object degenerates near a boundary.

The shelf ladder

F05 also studies a second ladder of means, built from chart functions rather than from the HyperCore elevator:

S_1 = AM,
S_2 = GM,
S_3 = M_3,
S_4 = M_att.

Here M_3 is the storage mean of commutative power Cpow, generated by ln(x*ln(x)), while M_att is the quasi-arithmetic attractor mean generated through a regular superlogarithm slog_B.

The shelf ladder agrees with the HyperCore’s storage means through rank three. At rank four the constructions separate. The HyperCore’s M_4 is non-quasi-arithmetic and generated by its own nonseparable channel. The shelf mean M_att is quasi-arithmetic, depends on the selected base B, and presents rank four through a Symmetric-Core chart.

The shelf presentation at 7/2 is

ISHE-S = P_(M_3,M_att),

the Gauss compound of these two means. A related presentation compounds M_3 with the geometric mean transported through slog_B. Neither is the HyperCore’s base-free protocol ISHE-H.

Why the floor is blind

At the floor x=1, both

ln(x) and slog_B(x)

vanish simply. Their first-order behaviour differs only by a nonzero scale, and affine changes of a quasi-arithmetic generator leave its mean unchanged. Consequently, the shelf ladder’s first-order floor jets are

Fl_1[S_1] = Fl_1[S_2] = AM,

Fl_1[S_3] = Fl_1[S_4] = GM.

The shelf floor therefore realises the partition

{1,2}, {3,4}.

It cannot distinguish the logarithmic and superlogarithmic attractor means. Both shelf presentations at 7/2 have the bare geometric mean as their floor limit:

Fl_1[ISHE-S] = GM.

This differs from the HyperCore protocol ISHE-H, whose conditional floor is GH. The distinction demonstrates why a shelf presentation cannot substitute for the HyperCore’s own rank-four storage mean.

The second order

The first-order floor is blind, but the second-order term detects the curvature of the chart function.

Suppose

theta(1+delta) = c*delta+c_2*delta^2+...

and define the curvature ratio

r = c_2/c.

For an attractor mean generated by ln(x*theta(x)), F05 proves the universal expansion

M(1+epsilon*a,1+epsilon*b)

= 1+epsilon*sqrt(a*b)

+ epsilon^2*((1+r)/2)*sqrt(a*b)*(sqrt(a)-sqrt(b))^2

+ O(epsilon^3).

The correction always has the same operand-dependent shape,

sqrt(a*b)*(sqrt(a)-sqrt(b))^2.

Only its scalar coefficient remembers the chart.

For theta=ln, the ratio is

r_3 = -1/2.

For the regular superlogarithm at base B=sqrt(2), the corrected value is

r_4 = 0.70351204019672301403....

If two compound legs have the same first-order jet, their second-order terms are averaged by the compound. Therefore,

ISHE-S(1+epsilon*a,1+epsilon*b)

= 1+epsilon*sqrt(a*b)

+ epsilon^2*((2+r_3+r_4)/4)

*sqrt(a*b)*(sqrt(a)-sqrt(b))^2

+ O(epsilon^3).

At B=sqrt(2), the scalar coefficient is

0.55087801004918075351....

The alternative slog-chart protocol has coefficient

(1+r_3+r_4)/4

= 0.30087801004918075351....

High-precision calculations confirm these formulas through the expected third-order remainder.

The second-order jet therefore distinguishes the logarithmic and superlogarithmic legs through one number, r_4. It recovers the local curvature of the shelf chart at the floor. It does not recover the HyperCore’s rank-four mean, which belongs to a different construction and already has a different first-order floor jet.

The ceiling

The superlogarithmic shelf has a second distinguished boundary. For

1<B<exp(1/e),

let L_B be the attracting fixed point of x->B^x, with multiplier

Lambda_B = ln(L_B).

As delta->0+, Koenigs’s coordinate gives

ln(1/Lambda_B)*slog_B(L_B-delta)

= ln(1/delta)+k_B+O(delta),

where

k_B = ln|psi_B(1)|

depends on the chosen base.

Thus the superlogarithm diverges logarithmically at the ceiling. After writing

x = L_B-epsilon*z,

its natural chart becomes

u = Lambda-ln(z),

with

Lambda = ln(1/epsilon)+k_B.

In this scaling, the rank-three leg M_3, which is regular at the ceiling, approaches the arithmetic mean. The superlogarithmic attractor leg approaches the geometric mean, but only at the slow logarithmic rate

1/ln(1/epsilon).

Their compound consequently approaches Gauss’s mean:

Ce_1[ISHE-S] = AGM.

More precisely,

ISHE-S(L_B-epsilon*a,L_B-epsilon*b)

= L_B-epsilon*[AGM(a,b)

+ C(a,b)/(ln(1/epsilon)+k_B)

+ O(ln^(-2)(1/epsilon))].

The correction C(a,b) is an explicit positive orbit sum over Gauss’s AGM iteration. It is symmetric, homogeneous of degree one, zero on the diagonal and positive away from it. At (a,b)=(1,2),

C(1,2) = 0.0434168034108973....

The base affects the expansion only through the additive constant k_B in the logarithmic scale. The leading ceiling mean AGM is independent of the base.

The approach at the floor and ceiling is therefore qualitatively different:

• at the floor, corrections occur in powers of epsilon;
• at the ceiling, corrections occur in powers of 1/ln(1/epsilon).

This difference reflects the analyticity of the chart functions near 1 and the logarithmic singularity of slog_B near L_B.

The Elevator Shadow

The first edition proposed that one rank casts a shadow on the next at the ceiling of the superlogarithmic shelf. F05 v1.0 resolves this claim in two senses.

At the level of first-order jets, the statement holds:

Ce_1[S_4] = Fl_1[S_3] = GM,

Ce_1[ISHE-S] = Fl_1[SHM] = AGM.

The shelf presentation at 7/2 therefore sees at its ceiling the same classical mean that the HyperCore’s rank-5/2 protocol sees at its floor.

For the full germs in the identity chart, the statement fails. The ceiling correction is positive and logarithmic:

C(a,b)/ln(1/epsilon).

The floor correction of SHM is algebraic in its scale and has coefficient D_desc(a,b), with

D_desc(1,2) = -0.0310397833....

The signs differ at (1,2), and the operand-dependent ratios are not proportional. No positive rescaling of the same operands can therefore identify the two germs to the next order.

The strong matched-chart version remains open: a nontrivial change of operand coordinates might conceivably relate the two families beyond first order. The identity chart is excluded.

F05 clarifies what operational descent can and cannot mean. The HyperCore elevator is a global generative construction; a boundary jet is a local, forgetful limit. The shelf presentation is indistinguishable from the logarithmic rung at first order near the floor, but its superlogarithmic curvature appears at second order. At the ceiling, the same presentation contracts to Gauss’s AGM with an explicit logarithmic correction. The Elevator Shadow is therefore exact for first-order jets and false for the complete germs in the natural identity chart. The monograph contains no physics and bears on no famous open problem. Its contribution is a precise boundary calculus for distinguishing the HyperCore’s own hierarchy from chart-defined shelf presentations.

F papers belong to the M43 series.