PAIR A
PANTS MATH / FOUNDATIONS & APPLICATIONSFIELD NOTES № 001 / 2026
A new math
01 — THE FOUNDATIONAL EQUATION
Equal and opposite shaping cancels: (ℝ³, +) is an Abelian group.
View the axioms ↓01 / THE EQUATION LAB
Silhouette algebra.
Pick two silhouettes. Add their shaping.
Change the inputs. The rules hold.
0 · no pockets30 cm · deep utility
PAIR B
PAIR C
YOUR SOLUTION↙ Q.E.D.
Straight cut
ON THE CHALKBOARD
RESULT APPEARANCE
Color intensity and shape rules
Intensity ranges from neutral grey (0%) to the selected solid color (100%). Each color is represented by a signed RGB deviation from grey, scaled by intensity. Add and subtract operate on these values; blend and dress combination take their mean. Screen colors are clipped only when displayed.
The shape slider selects a repeating fabric motif: plain, circle, square, triangle or diamond. Motifs have signed counts: addition combines them, subtraction cancels matching shapes, and averaging reduces their weights. Negative motifs have dashed outlines. These appearance coordinates extend the silhouette model without altering its geometric axioms.
EXPORT CURRENT RESULT
128 × 128 px · transparent background
Coordinates are (flare, volume, rise). These are style parameters, not body measurements. We add changes from one shared baseline, not physical pairs of trousers.
02 / ALGEBRAIC FOUNDATIONS
Pants Math Axioms
A three-dimensional real vector space
with signed silhouette coordinates
and explicitly defined operations.
FIRST, DEFINE A PAIR
P = (f, v, r)
Every formal pair lives in ℝ³. Three signed coordinates describe a deviation from straight, regular-volume, regular-rise pants.
f tapered ← 0 → flaredv slim ← 0 → wider low rise ← 0 → high rise
The basis pairs are F = (1, 0, 0), V = (0, 1, 0), H = (0, 0, 1). Therefore P = fF + vV + rH. Fabric, pockets, gender and price are outside this base vector model. The dress operation adds a separate pocket-depth parameter. Same coordinates = same mathematical pair.
01Closure P + Q ∈ ℝ³
Add corresponding coordinates: (f₁ + f₂, v₁ + v₂, r₁ + r₂). Any two finite formal pairs produce another formal pair. Wearability is a separate constraint.
02Commutativity P + Q = Q + P
Tapered + flared is the same as flared + tapered. The order of addition cannot change the outfit.
03Associativity (P + Q) + R = P + (Q + R)
Regroup any finite sum. The silhouette stays the same. Turn on the third pair in the equation lab to try it.
04The zero pair P + S = P
S = (0, 0, 0): straight, regular-volume, regular-rise. Zero denotes the neutral shaping baseline.
05Every pair has an anti-pair P + (−P) = S
−P = (−f, −v, −r). A high-rise wide flare has a low-rise slim taper as its formal inverse. Signed shape, not negative cloth.
06Scalar identity & composition 1P = P; a(bP) = (ab)P
For real a and b, aP = (af, av, ar). Scaling changes the intensity of the deviation. 0P = S; 2F is double flare intensity, not twice the hem width.
07Distribute over pairs a(P + Q) = aP + aQ
Amplifying the combined silhouette is equivalent to amplifying each input and then adding.
08Distribute over scalars (a + b)P = aP + bP
Three portions of a shaping direction plus two more make five. Division by a nonzero scalar works; division by the zero pair is not defined.
Open the pattern: how coordinates become a drawing +
For normalized distance z from crotch (0) to hem (1), set leg width wₚ(z) = w₀ exp(0.18v + 0.30fz) and rise hₚ = h₀ exp(0.12r). Width and rise remain positive for every finite formal coordinate.
Adding coordinates multiplies relative width profiles: wₚ₊Q(z) / w₀ = [wₚ(z) / w₀][wQ(z) / w₀]. That is why equal taper and flare cancel. The foundational equation assumes matched intensities and the same baseline volume and rise.
The neutral pair is a chosen reference, not a universal ideal of fit. Drawings are schematic: coordinates beyond ±3 are visually capped and labeled, while all calculations retain the original values. Mathematically ℝ³ is unbounded; this browser computes finite-precision approximations.
03 / INVERSES & LIMITSNegative values.
Negative values.
Infinite sequences.
Negative coordinates define additive inverses.
Infinite sequences describe convergence
and unbounded growth.
The additive
inverse.
+=
Taper / slim / lowFlare / wide / high
All three coordinates reverse. Adding a pair to its inverse cancels its shaping and yields the neutral straight pair.
Convergence
and divergence.
∞ − ∞ is not a straight pair.
nF − nF = S, but 2nF − nF = nF. Both look like “∞ − ∞” if you throw away the sequence. Keep the sequence. Infinity is not a member of ℝ³, so it is not an ordinary addable pair.
DISTANCE
d(P,Q) = ‖P − Q‖₂Euclidean distance quantifies separation between two silhouettes in the defined coordinate space.
CONTINUOUS TRANSITIONS
P(t) = (1 − t)P + tQFor 0 ≤ t ≤ 1, walk smoothly between two looks. For t outside that range, extrapolate.
FASHION VELOCITY
P′(t) = (f′, v′, r′)The rate at which a silhouette changes. Trends become paths; revivals can form loops.
04 / FASHION, HISTORY & SCIENCEThe historical development
The historical development
of pants.
Clothing is technology. It also carries power,
identity, labor and belonging. Our three
coordinates capture only the silhouette.
A few selected moments, not a single origin story or an inevitable march of progress. Historical artifacts are evidence; the Pants Math interpretations are ours.
05 / MATRIX TRANSFORMATIONSPants are nouns.
Pants are nouns.
Operators are verbs.
Add silhouettes in any order.
Transform them in a different order,
and the result can change.
Try the default experiment. Start with a wide pair V. Feed its volume into flare with C, then double that flare with D. Reverse the order. The resulting silhouettes differ.
FIRST / A
THEN / B
The operational grammar: compose, translate, remember +
01 / Expressions have types
P ::= S | F | V | H | (P + P) | aP | M(P)
P is a pair, a is a real number, M is a 3 × 3 real matrix. Parentheses bind first. Read M(P) as “apply M to P.” Multiplication of two pants is deliberately undefined.
02 / Transformations compose
B(A(P)) = (BA)P
Read compositions from right to left. Matrix composition is associative and generally not commutative. Linear operators distribute over pants addition. Only invertible matrices can undo every transformation; erasing flare loses information.
03 / Affine offsets
P′ = MP + b
An affine operation can add a fixed rise or flare. With homogeneous coordinates (f, v, r, 1), use the 4 × 4 block matrix [M b; 0 1]. Unlike a linear map, an affine map need not preserve S.
04 / Turn style into a sequence
Pₜ₊₁ = MPₜ + bₜ
Now pants have a state and a next state. Add a learned preference score to choose the next move. Feedback gives the system adaptation; a matrix alone does not give it a mind.
06 / PANTS AII think,
I think,
therefore I hem.
Train a preference model through paired comparisons.
Observe how each choice changes its parameters.
WHICH WOULD YOU WEAR?LESSON 01
OR
Pants AI: learning, agency and limitations +
What really learns
A small online preference model updates three weights w after every choice. Its score is w·P. The chance assigned to choosing A over B is σ(w·(A−B)), where σ(x) = 1/(1+e⁻ˣ).
For chosen C and rejected R, the update is w ← 0.995w + 0.22[1−σ(w·(C−R))](C−R). The decay gently discounts old evidence. “Neither” skips without training.
How it proposes a look
The proposal has coordinates P* = 2 tanh(w), component by component. This turns each learned direction into a bounded silhouette. The meters show weights, not confidence or personality traits. The percentages are model predictions, not calibrated certainty.
Learning and agency
State, memory, feedback and action enable adaptive behavior. This implementation is a local preference-learning model. Adaptation alone does not establish consciousness or sentience.
Future research directions
A future system might model context, ask when uncertain, sense movement, and propose transformations within wearer-defined limits. Smart textiles could give it a physical output. Such a system would require explicit wearer control over its objectives and permitted transformations.
Develop the next
Pants Math theorem.
Start with a pair. Define your reference.
Make a transformation. See what holds.