{"id":1022,"date":"2026-06-15T13:51:50","date_gmt":"2026-06-15T13:51:50","guid":{"rendered":"https:\/\/tensorzen.online\/?p=1022"},"modified":"2026-06-15T13:51:50","modified_gmt":"2026-06-15T13:51:50","slug":"the-determinant-volume-collapse-and-singular-matrices","status":"publish","type":"post","link":"https:\/\/tensorzen.online\/?p=1022","title":{"rendered":"The Determinant: Volume, Collapse, and Singular Matrices"},"content":{"rendered":"\n
This notebook builds the geometric and algebraic understanding of the determinant from first principles, culminating in why $\\det(A) = 0$ is the exact condition for singularity.<\/p>\n\n\n\n
import numpy as np\nimport matplotlib.pyplot as plt\nimport matplotlib.patches as mpatches\nfrom matplotlib.patches import Polygon\nfrom matplotlib.collections import PatchCollection\n\ndef style_ax(ax, title='', xlim=(-4, 4), ylim=(-4, 4)):\n ax.set_xlim(*xlim); ax.set_ylim(*ylim)\n ax.set_aspect('equal')\n ax.axhline(0, color='gray', lw=0.7)\n ax.axvline(0, color='gray', lw=0.7)\n ax.set_title(title, fontsize=11, pad=8)\n ax.set_xlabel('$x_1$'); ax.set_ylabel('$x_2$')\n ax.grid(True, linestyle='--', alpha=0.3)\n\ndef draw_arrow(ax, vec, origin=None, color='black', lw=2, label=None, offset=None):\n o = np.zeros(2) if origin is None else np.array(origin)\n ax.annotate('', xy=o + vec, xytext=o,\n arrowprops=dict(arrowstyle='->', color=color, lw=lw))\n if label:\n pos = o + vec + (offset if offset is not None else np.array([0.1, 0.1]))\n ax.text(*pos, label, color=color, fontsize=12, fontweight='bold')\n\ndef draw_parallelogram(ax, c1, c2, color='steelblue', alpha=0.25, edge_lw=1.5):\n """Draw the parallelogram spanned by two column vectors."""\n corners = np.array([[0,0], c1, c1+c2, c2, [0,0]])\n ax.fill(corners[:,0], corners[:,1], color=color, alpha=alpha)\n ax.plot(corners[:,0], corners[:,1], color=color, lw=edge_lw)<\/code><\/pre>\n\n\n\n
1. The Determinant as a Volume Scaling Factor<\/h2>\n\n\n\n
The most fundamental interpretation of the determinant is geometric, not algebraic:<\/p>\n\n\n\n
\n
$\\det(A)$ is the signed volume of the parallelepiped formed by the columns of $A$.<\/strong><\/p>\n<\/blockquote>\n\n\n\n
In 2D, this is the signed area of the parallelogram spanned by the two column vectors $a_1$ and $a_2$:<\/p>\n\n\n\n
The determinant is a uniform volume scaling factor<\/strong> \u2014 it tells you by how much the transformation stretches or compresses all of space.<\/p>\n\n\n\n
A negative determinant means the transformation includes a reflection \u2014 the columns have swapped their relative handedness compared to the standard basis.<\/p>\n\n\n\n
3. Why $\\det(A) = 0$ Means Singular: Dimension Collapse<\/h2>\n\n\n\n
When $\\det(A) = 0$, the transformation maps $\\mathbb{R}^n$ into a strictly lower-dimensional subspace<\/strong>. In 2D, the entire plane gets flattened onto a line (or a point). The parallelogram has zero area because the two column vectors are collinear<\/strong> \u2014 they both point along the same line.<\/p>\n\n\n\n
$$\\det(A) = 0 \\iff \\text{columns of } A \\text{ are linearly dependent} \\iff \\text{image of } A \\text{ has dimension} < n$$<\/p>\n\n\n\n
# Singular matrix: col2 = 2 * col1\nA_sing = np.array([[1.0, 2.0],\n [0.5, 1.0]])\n\n# A grid of input points\ngrid_pts = np.array([[x, y] for x in np.linspace(-2,2,9)\n for y in np.linspace(-2,2,9)]).T\ntransformed = A_sing @ grid_pts\n\nfig, axes = plt.subplots(1, 2, figsize=(13, 5))\n\n# Input space\nax = axes[0]\nax.scatter(grid_pts[0], grid_pts[1], color='steelblue', s=30, zorder=3)\ndraw_arrow(ax, A_sing[:,0], color='crimson', lw=2.5, label='$a_1$', offset=np.array([ 0.1,-0.25]))\ndraw_arrow(ax, A_sing[:,1], color='darkorange', lw=2.5, label='$a_2 = 2a_1$', offset=np.array([0.1, 0.1]))\nstyle_ax(ax, title='Input: 2D grid of points', xlim=(-3,3), ylim=(-3,3))\n\n# Output space \u2014 everything collapses to a line\nax = axes[1]\nax.scatter(transformed[0], transformed[1], color='crimson', s=30, zorder=3, label='Transformed points')\n# draw the line that the space collapsed onto\nt = np.linspace(-4, 4, 100)\ncol1 = A_sing[:, 0]\nax.plot(col1[0]*t, col1[1]*t, 'k--', lw=1.5, alpha=0.5, label='Image: 1D line (col space)')\nax.legend(fontsize=9)\nstyle_ax(ax, title=f'Output: entire 2D plane collapses to a 1D line\\n$\\\\det(A) = {np.linalg.det(A_sing):.4f} \\\\approx 0$',\n xlim=(-4,4), ylim=(-3,3))\n\nfig.suptitle('Singular matrix: $a_2 = 2a_1$ \u2014 columns are linearly dependent', fontsize=13)\nplt.tight_layout()\nplt.show()\n\nprint(f"A_sing =\\n{A_sing}")\nprint(f"det(A_sing) = {np.linalg.det(A_sing):.6f}")\nprint(f"rank(A_sing) = {np.linalg.matrix_rank(A_sing)} (collapsed from 2 to 1)")<\/code><\/pre>\n\n\n\n
<\/figure>\n\n\n\n
A_sing =\n[[1. 2. ]\n [0.5 1. ]]\ndet(A_sing) = 0.000000\nrank(A_sing) = 1 (collapsed from 2 to 1)<\/pre>\n\n\n\n
4. The Mechanism: Alternating Multilinearity<\/h2>\n\n\n\n
Why does linear dependence force $\\det = 0$? The answer lies in two algebraic properties that define the determinant:<\/p>\n\n\n\n
Multilinear<\/strong> \u2014 scaling any one column scales the determinant by the same factor: $$\\det([\\ldots, \\alpha a_j, \\ldots]) = \\alpha \\cdot \\det([\\ldots, a_j, \\ldots])$$<\/p>\n\n\n\n
Alternating<\/strong> \u2014 swapping any two columns negates the determinant, and a repeated column gives zero: $$\\det([\\ldots, a_i, \\ldots, a_i, \\ldots]) = 0$$<\/p>\n\n\n\n
Now suppose column $k$ is a linear combination of the others: $a_k = \\sum_{j \\neq k} c_j a_j$. By multilinearity, the determinant expands into a sum of terms \u2014 each one containing a repeated column. The alternating property kills each term. Therefore:<\/p>\n\n\n\n
A matrix $A$ is invertible iff $Ax = b$ has a unique solution for every $b$. This fails when $A$ collapses a dimension \u2014 either outputs are unreachable, or multiple inputs share the same output.<\/p>\n\n\n\n
The product rule<\/strong> $\\det(AB) = \\det(A)\\det(B)$ makes the impossibility of inverting a singular matrix algebraically transparent:<\/p>\n\n\n\n
This gives another route to the same conclusion: $\\det(A) = 0$ iff at least one $\\lambda_i = 0$. A zero eigenvalue means some nonzero vector $v$ satisfies $Av = 0 \\cdot v = 0$ \u2014 it gets annihilated by the transformation. That is exactly a nontrivial null space.<\/p>\n\n\n\n
And since singular values $\\sigma_i = \\sqrt{\\lambda_i(A^\\top A)}$, a zero singular value is the same condition expressed through the SVD lens.<\/p>\n\n\n\n
for M, name in [(A_inv, 'Non-singular'), (A_sing, 'Singular')]:\n eigvals = np.linalg.eigvals(M)\n singular_vals = np.linalg.svd(M, compute_uv=False)\n det_direct = np.linalg.det(M)\n det_via_eigs = np.prod(eigvals).real\n\n print(f"=== {name} ===")\n print(f" Eigenvalues : {np.round(eigvals.real, 4)}")\n print(f" det (direct) : {det_direct:.6f}")\n print(f" det (\u220f \u03bb\u1d62) : {det_via_eigs:.6f} \u2713")\n print(f" Singular values: {np.round(singular_vals, 6)}")\n print(f" Rank : {np.linalg.matrix_rank(M)}")\n print()<\/code><\/pre>\n\n\n\n
======================================================= Non-singular A ======================================================= det(A) : 2.750000 Rank = n (2)? : True (rank = 2) All eigenvalues \u2260 0? : True [2.309 1.191] All singular values > 0? : True [2.309017 1.190983] Null space trivial ({0})? : True Invertible? : True
======================================================= Singular A ======================================================= det(A) : 0.000000 Rank = n (2)? : False (rank = 1) All eigenvalues \u2260 0? : False [2. 0.] All singular values > 0? : False [2.5 0. ] Null space trivial ({0})? : False Invertible? : False<\/pre>\n\n\n\n
Summary<\/h2>\n\n\n\n
Condition<\/th>
What it says<\/th><\/tr><\/thead>
$\\det(A) = 0$<\/td>
Transformation destroys volume \u2014 space is collapsed<\/td><\/tr>
Columns linearly dependent<\/td>
At least one column lies in the span of the others<\/td><\/tr>
$\\text{rank}(A) < n$<\/td>
Image has lower dimension than domain<\/td><\/tr>
$Ax = 0$ has nontrivial solutions<\/td>
Some nonzero vector is annihilated<\/td><\/tr>
$A$ is not invertible<\/td>
The transformation cannot be undone<\/td><\/tr>
At least one eigenvalue $= 0$<\/td>
Some direction gets completely flattened<\/td><\/tr>
At least one singular value $= 0$<\/td>
Same condition via SVD<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
These are not separate facts. They are one geometric fact<\/strong> \u2014 the transformation collapses at least one dimension \u2014 expressed from different mathematical viewpoints.<\/p>\n\n\n\n
The determinant is the single scalar that captures whether this collapse has occurred: nonzero means the full-dimensional structure of space is preserved and the transformation is reversible; zero means information has been irretrievably lost.<\/p>\n\n\n\n
8. What Does “Spanned By” Mean?<\/h2>\n\n\n\n
The phrase “spanned by”<\/strong> appeared throughout this notebook (e.g. “the parallelogram spanned by $a_1$ and $a_2$”<\/em>). Here is what it means precisely.<\/p>\n\n\n\n
Definition<\/h3>\n\n\n\n
The span<\/strong> of vectors $v_1, v_2, \\ldots, v_k$ is the set of every vector reachable by scaling and adding them in any combination:<\/p>\n\n\n\n
“The space spanned by $v_1$ and $v_2$” means this entire set \u2014 all possible linear combinations.<\/p>\n\n\n\n
What spans look like geometrically<\/h3>\n\n\n\n
Vectors<\/th>
Span<\/th><\/tr><\/thead>
One nonzero vector in $\\mathbb{R}^2$<\/td>
A line through the origin<\/td><\/tr>
Two independent vectors in $\\mathbb{R}^2$<\/td>
The entire plane $\\mathbb{R}^2$<\/td><\/tr>
Two dependent<\/strong> vectors in $\\mathbb{R}^2$<\/td>
Only a line (same as one vector)<\/td><\/tr>
Three independent vectors in $\\mathbb{R}^3$<\/td>
All of $\\mathbb{R}^3$<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n
Key insight:<\/strong> linear dependence shrinks the span. If one vector is already a combination of the others, adding it contributes nothing new.<\/p>\n\n\n\n
Connection to this notebook<\/h3>\n\n\n\n
\n
“Parallelogram spanned by $a_1$ and $a_2$”<\/em> \u2014 the filled shape formed by all $\\alpha_1 a_1 + \\alpha_2 a_2$ with $0 \\leq \\alpha_1, \\alpha_2 \\leq 1$<\/li>\n\n\n\n
“Column space”<\/em> \u2014 another name for the span of the columns of $A$: all vectors $Ax$ can possibly reach<\/li>\n\n\n\n
“$a_k \\in \\text{span}(\\text{other columns})$”<\/em> \u2014 column $k$ is redundant; it adds no new direction, so the determinant is zero<\/li>\n<\/ul>\n\n\n\n
fig, axes = plt.subplots(1, 3, figsize=(16, 5))\n\n# \u2500\u2500 Case 1: span of one vector = a line \u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\nax = axes[0]\nv1 = np.array([1.5, 0.8])\nt = np.linspace(-2.5, 2.5, 200)\nspan_pts = np.outer(t, v1) # all \u03b1\u00b7v1\nax.plot(span_pts[:, 0], span_pts[:, 1], color='steelblue', lw=2, label='span($v_1$) = line')\n# sample a few representative points\nfor alpha in [-2, -1, 0, 1, 2]:\n pt = alpha * v1\n ax.scatter(*pt, color='steelblue', s=40, zorder=4)\ndraw_arrow(ax, v1, color='crimson', lw=2.5, label='$v_1$', offset=np.array([0.05, 0.12]))\nax.legend(fontsize=9)\nstyle_ax(ax, title='span($v_1$): one vector \u2192 a line\\nthrough the origin', xlim=(-3,3), ylim=(-2.5,2.5))\n\n# \u2500\u2500 Case 2: span of two independent vectors = the plane \u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\nax = axes[1]\nv1 = np.array([1.5, 0.5])\nv2 = np.array([0.3, 1.5])\n# shade the reachable region (the whole plane \u2014 approximate with a dense grid)\nalphas = np.linspace(-2, 2, 18)\npts = np.array([a1*v1 + a2*v2 for a1 in alphas for a2 in alphas])\nax.scatter(pts[:, 0], pts[:, 1], color='steelblue', s=15, alpha=0.4, label='sample of span($v_1, v_2$)')\ndraw_arrow(ax, v1, color='crimson', lw=2.5, label='$v_1$', offset=np.array([ 0.05,-0.25]))\ndraw_arrow(ax, v2, color='darkorange', lw=2.5, label='$v_2$', offset=np.array([-0.4, 0.1]))\n# highlight the parallelogram (\u03b1 \u2208 [0,1])\ndraw_parallelogram(ax, v1, v2, color='steelblue', alpha=0.3)\nax.legend(fontsize=9)\nstyle_ax(ax, title='span($v_1, v_2$): two independent vectors\\n\u2192 fills the entire plane',\n xlim=(-3.5, 3.5), ylim=(-3, 3.5))\n\n# \u2500\u2500 Case 3: span of two dependent vectors = still just a line \u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\u2500\nax = axes[2]\nv1 = np.array([1.2, 0.6])\nv2 = 2.0 * v1 # v2 is just a scaled v1\nt = np.linspace(-2, 2, 200)\nspan_pts = np.outer(t, v1)\nax.plot(span_pts[:, 0], span_pts[:, 1], color='steelblue', lw=2, label='span($v_1, v_2$) = still a line')\ndraw_arrow(ax, v1, color='crimson', lw=2.5, label='$v_1$', offset=np.array([ 0.05,-0.25]))\ndraw_arrow(ax, v2, color='darkorange', lw=2.5, label='$v_2 = 2v_1$', offset=np.array([ 0.1, 0.1]))\nax.legend(fontsize=9)\nstyle_ax(ax, title='span($v_1, v_2$): two dependent vectors\\n\u2192 still only a line ($v_2$ adds nothing)',\n xlim=(-3, 3), ylim=(-2, 2.5))\n\nfig.suptitle('"Spanned by": the set of all linear combinations of a collection of vectors', fontsize=13)\nplt.tight_layout()\nplt.show()<\/code><\/pre>\n\n\n\n<\/figure>\n","protected":false},"excerpt":{"rendered":"
This notebook builds the geometric and algebraic understanding of the determinant from first principles, culminating in why $\\det(A) = 0$ is the exact condition for singularity. 1. The […]<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[25,6],"tags":[],"class_list":["post-1022","post","type-post","status-publish","format-standard","hentry","category-in-english","category-matchematics"],"_links":{"self":[{"href":"https:\/\/tensorzen.online\/index.php?rest_route=\/wp\/v2\/posts\/1022","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/tensorzen.online\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/tensorzen.online\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/tensorzen.online\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/tensorzen.online\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=1022"}],"version-history":[{"count":1,"href":"https:\/\/tensorzen.online\/index.php?rest_route=\/wp\/v2\/posts\/1022\/revisions"}],"predecessor-version":[{"id":1029,"href":"https:\/\/tensorzen.online\/index.php?rest_route=\/wp\/v2\/posts\/1022\/revisions\/1029"}],"wp:attachment":[{"href":"https:\/\/tensorzen.online\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=1022"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/tensorzen.online\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=1022"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/tensorzen.online\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=1022"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
![\"\"](\"https:\/\/tensorzen.online\/wp-content\/uploads\/2026\/06\/image-14-1024x322.png\")
<\/figure>\n\n\n\n![\"\"](\"https:\/\/tensorzen.online\/wp-content\/uploads\/2026\/06\/image-15-1024x358.png\")
<\/figure>\n\n\n\n![\"\"](\"https:\/\/tensorzen.online\/wp-content\/uploads\/2026\/06\/image-16-1024x435.png\")
<\/figure>\n\n\n\n![\"\"](\"https:\/\/tensorzen.online\/wp-content\/uploads\/2026\/06\/image-17-1024x332.png\")
<\/figure>\n\n\n\n![\"\"](\"https:\/\/tensorzen.online\/wp-content\/uploads\/2026\/06\/image-18-1024x481.png\")
<\/figure>\n\n\n\n![\"\"](\"https:\/\/tensorzen.online\/wp-content\/uploads\/2026\/06\/image-19-1024x334.png\")
<\/figure>\n","protected":false},"excerpt":{"rendered":"