00 / 先给答案/ answer first

先给结论：超曲面，就是比周围空间少一个维度Begin with the conclusion: a hypersurface has one dimension fewer than its ambient space

测试版 / 未完成beta / unfinished交互、措辞与数学表达仍在复核。Interaction, wording, and mathematical communication remain under review.

在光滑情形下，你可以把它理解为一种In the smooth case, understand it as a geometric object of余维为 1codimension one的几何对象：二维空间里的一条曲线、三维空间里的一个曲面，以及四维空间里的三维球面，服从同一条规律。: a curve in two-dimensional space, a surface in three-dimensional space, and a three-sphere in four-dimensional space all follow the same rule.

Mⁿ⁻¹ ⊂ ℝⁿ, codim M = 1

二维空间Two-dimensional space曲线Curve $S 1$

三维空间Three-dimensional space球面Sphere $S 2$

四维空间Four-dimensional space三维球面Three-sphere $S 3$

01 / 三个问题/ three questions

理解超曲面，只需回答三个问题To understand a hypersurface, answer three questions

后面的每一屏只做一件事，并且都回到这三条主线。先建立维数直觉，再解释约束，最后才处理看不见的第四维。Each screen that follows does one thing and returns to these three lines: build intuition for dimension, explain constraints, and only then address an unseen fourth dimension.

维数是什么？What is dimension?

数一数：留在对象上时，有多少个彼此独立的局部方向。Count the independent local directions available while remaining on the object.

为什么一个方程会少一维？Why does one equation remove one dimension?

一个独立约束通常会拿走一个自由度。One independent constraint usually removes one degree of freedom.

四维对象怎么看？How can a four-dimensional object be viewed?

用切片逐层观察，再用投影同时压入三维。Inspect it layer by layer through slices, then compress all layers into three dimensions by projection.

02 / 零维/ zero dimensions

0D 没有可走的方向0D has no direction in which to move

紫色影子记录任意输入；亮点始终被规则压回原位。只要必须留在这个对象上，就不存在任何非零的局部运动方向。The purple shadow records arbitrary input while the bright point is always constrained back to its position. If it must remain on the object, no nonzero local direction of motion exists.

dim = 0

03 / 一个自由度/ one degree of freedom

1D 只有一种独立走法1D has one independent direction of motion

刚才的点沿The previous point extends along the $u$ 方向展开成一条线。任意偏离直线的输入都会被投影回来；左右只是相反朝向，不是两个独立维度。direction to form a line. Any input off the line is projected back; left and right are opposite orientations, not two independent dimensions.

P₁(u) = (u, 0, 0), dim = 1

04 / 两个自由度/ two degrees of freedom

2D 需要第二个、不可替代的方向2D requires a second, independent direction

整条线沿新的The entire line sweeps along a new $v$ 方向扫过，留下一个面。现在用direction, leaving a plane. Combinations of $u$ 与and $v$ 的组合，可以到达面内任意附近位置；但第三个独立方向仍然不存在。can now reach any nearby point in the plane, but a third independent direction still does not exist.

P₂(u, v) = (u, v, 0), dim = 2

05 / 三个自由度/ three degrees of freedom

3D 的关键不是变厚，而是出现第三个独立方向The key to 3D is not thickness but a third independent direction

平面沿The plane sweeps along the $z$ 方向扫过，形成体积。画面把direction to form a volume. The image separates $u, v$ 与第三坐标from the third coordinate $z$ 分开，高亮点和截面因此同步穿过整个体积。so the highlighted point and cross-section pass through the volume together.

P₃(u, v, z) = (u, v, z), dim = 3

06 / 重复出现的规律/ a recurring pattern

边界每次少一维，A boundary loses one dimension each time, $S 3$ 因而自然出现so the pattern appears naturally

线段的边界是两个点，圆盘的边界是圆，实心球的边界是球面。模式继续一层：四维球体的边界，就是三维球面A line segment is bounded by two points, a disk by a circle, and a solid ball by a sphere. One dimension higher, the boundary of a four-ball is a three-sphere $S 3$ 。

∂B¹ = S⁰, ∂B² = S¹, ∂B³ = S², ∂B⁴ = S³

这只是建立直觉的一条路径。一般的超曲面不必都以“球体边界”的形式出现。This is only one route to intuition. A general hypersurface need not arise as the boundary of a ball.

07 / 一个独立约束/ one independent constraint

一个独立约束，通常减少一个自由度One independent constraint usually removes one degree of freedom

在所有候选位置中加入一个方程：数轴留下两个点，平面留下一个圆，三维空间留下球面。再加入第四个坐标 $w$，规律不变。Impose one equation on all candidate positions: the number line leaves two points, the plane a circle, and three-dimensional space a sphere. Add a fourth coordinate $w$, and the rule remains unchanged.

x² = 1 ⇒ S⁰

x² + y² = 1 ⇒ S¹

x² + y² + z² = 1 ⇒ S²

x² + y² + z² + w² = 1 ⇒ S³

更严格地说，在More precisely, near a regular point where $F (p) = 0$ 且and $\nabla F (p) \neq 0$ 的正则点附近，水平集, the level set $F = 0$ 局部具有locally has $n - 1$ 个自由度。degrees of freedom.

08 / 将问题降低一维/ lower the problem by one dimension

看不见高维对象时，先看切片When a higher-dimensional object cannot be seen, begin with slices

让一个三维单位球穿过二维平面。二维观察者从未看见整个球，只经历一连串截面：一点、小圆、大圆、小圆、一点。切片就是高维对象与低维世界的真实交集。Pass a three-dimensional unit sphere through a two-dimensional plane. A two-dimensional observer never sees the whole sphere, only a sequence of cross-sections: point, small circle, large circle, small circle, point. A slice is the actual intersection of a higher-dimensional object with a lower-dimensional world.

z = c ⟹ x² + y² = 1 − c², r(c) = √(1 − c²)

09 / 同一逻辑，提高一维/ the same logic, one dimension higher

固定Fixing $w$ ， $S 3$ 的切片就是普通的produces an ordinary slice $S 2$

三维球面Three-sphere $S 3$ 位于四维空间中。固定第四坐标in four-dimensional space. Fixing the fourth coordinate $w = c$ 后，只剩三个坐标，于是得到一个半径随leaves three coordinates and produces an ordinary sphere whose radius varies with $c$ 变化的普通球面。.

x² + y² + z² + w² = 1

w = c ⇒ x² + y² + z² = 1 − c²

S³ ∩ {w = c} ≅ S²_{√(1 − c²)}

10 / 切片不是投影/ slice is not projection

切片一次看一个A slice sees one $w$ ；投影把所有at a time; projection compresses all $w$ 同时压进三维into three dimensions at once

现在才看完整的Only now consider the complete $S 3$ 三维投影。数千个四维点先在涉及第四坐标的平面中旋转，再被压入三维。颜色编码three-dimensional projection. Thousands of four-dimensional points first rotate in planes involving the fourth coordinate and are then compressed into 3D. Color encodes $w$ ；你看到的是影子，不是对象本身。; what you see is a shadow, not the object itself.

R_xw(α), R_yw(β), R_zw(γ)

11 / 结论/ conclusion

因此，超曲面的核心只有一句话：余维为 1The central statement is therefore simple: a hypersurface has codimension one

前面的自由度、边界、约束、切片与投影，最终都回到同一个结论：在Degrees of freedom, boundaries, constraints, slices, and projections all return to one conclusion: in an $n$ 维环境空间中，一张光滑超曲面局部具有-dimensional ambient space, a smooth hypersurface locally has $n - 1$ 个自由度。degrees of freedom.