<?xml version="1.0" encoding="UTF-8" standalone="yes"?><oembed><version><![CDATA[1.0]]></version><provider_name><![CDATA[Azimuth]]></provider_name><provider_url><![CDATA[https://johncarlosbaez.wordpress.com]]></provider_url><author_name><![CDATA[John Baez]]></author_name><author_url><![CDATA[https://johncarlosbaez.wordpress.com/author/johncarlosbaez/]]></author_url><title><![CDATA[Symmetry and the Fourth Dimension (Part&nbsp;5)]]></title><type><![CDATA[link]]></type><html><![CDATA[<p><a href="https://johncarlosbaez.wordpress.com/2012/07/26/symmetry-and-the-fourth-dimension-part-4/">Last time</a> we saw that Platonic solids come in dual pairs, with the tetrahedron being dual to itself:  </p>
<div align="center"><a href="http://en.wikipedia.org/wiki/Tetrahedron"><img width="100" src="https://i0.wp.com/math.ucr.edu/home/baez/mathematical/tetrahedron_spinning.gif" alt="" /></a>
</div>
<div align="center"><a href="http://en.wikipedia.org/wiki/Cube"><img width="100" src="https://i2.wp.com/math.ucr.edu/home/baez/mathematical/cube_spinning.gif" alt="" /></a><a href="http://en.wikipedia.org/wiki/Octahedron"><img width="100" src="https://i0.wp.com/math.ucr.edu/home/baez/mathematical/octahedron_spinning.gif" alt="" /></a></div>
<div align="center"><a href="http://en.wikipedia.org/wiki/Dodecahedron"><img width="100" src="https://i1.wp.com/math.ucr.edu/home/baez/mathematical/dodecahedron_spinning.gif" alt="" /></a><a href="http://en.wikipedia.org/wiki/Icosahedron"><img width="100" src="https://i2.wp.com/math.ucr.edu/home/baez/mathematical/icosahedron_spinning.gif" alt="" /></a></div>
<p>When you have a dual pair, you can start chopping off the corners of one, more and more, and keep going until you reach the other.  Along the way you get some interesting shapes:</p>
<div align="center"><a href="http://en.wikipedia.org/wiki/Rectification_%28geometry%29"><img src="https://i1.wp.com/math.ucr.edu/home/baez/mathematical/450px-birectified_cube_sequence.jpg" /></a></div>
<p>At certain points along the way, we get <b><a href="http://en.wikipedia.org/wiki/Semiregular_polyhedron">semiregular polyhedra</a></b>, meaning that:</p>
<p>&bull; all the faces are regular polygons, and</p>
<p>&bull; there&#8217;s a symmetry carrying any corner to any other.  </p>
<p>Let&#8217;s see how it goes with the cube/octahedron pair.  And on the way, I&#8217;ll show you some diagrams that summarize what&#8217;s going on.  I&#8217;ll explain them later, but you can try to guess the pattern.  As a clue, I&#8217;ll say they&#8217;re based on the Coxeter diagram for the cube, which I explained last time:</p>
<div align="center">
<b> V&#8212;4&#8212;E&#8212;3&#8212;F </b>
</div>
<p>Here we go!</p>
<h3> Cube: <b> •&#8212;4&#8212;o&#8212;3&#8212;o </b></h3>
<p>First we have the <a href="http://en.wikipedia.org/wiki/Cube"><b>cube</b></a>, with all square faces:</p>
<div align="center"><a href="http://en.wikipedia.org/wiki/Cube"><br />
<img width="300" src="https://i2.wp.com/math.ucr.edu/home/baez/mathematical/400px-cube.png" /></a></div>
<h3> Truncated cube: <b> •&#8212;4&#8212;•&#8212;3&#8212;o </b></h3>
<p>Then we get the <a href="http://en.wikipedia.org/wiki/Truncated_cube"><b>truncated cube</b></a>, with octagons and triangles as faces:</p>
<div align="center"><a href="http://en.wikipedia.org/wiki/Truncated_cube"><br />
<img width="300" src="https://i2.wp.com/math.ucr.edu/home/baez/mathematical/400px-truncated_cube.png" /></a></div>
<h3> Cuboctahedron: <b> o&#8212;4&#8212;•&#8212;3&#8212;o </b></h3>
<p>Halfway through we get the aptly named <a href="http://en.wikipedia.org/wiki/Cuboctahedron"><b>cuboctahedron</b></a>, with squares and triangles as faces:</p>
<div align="center"><a href="http://en.wikipedia.org/wiki/Cuboctahedron"><br />
<img width="300" src="https://i2.wp.com/math.ucr.edu/home/baez/mathematical/400px-cuboctahedron.png" /></a></div>
<h3> Truncated octahedron: <b> o&#8212;4&#8212;•&#8212;3&#8212;• </b></h3>
<p>Then we get the <a href="http://en.wikipedia.org/wiki/Truncated_octahedron"><b>truncated octahedron</b></a>, with squares and hexagons as faces:</p>
<div align="center"><a href="http://en.wikipedia.org/wiki/Truncated_octahedron"><br />
<img width="300" src="https://i1.wp.com/math.ucr.edu/home/baez/mathematical/400px-truncated_octahedron.png" /></a></div>
<h3> Octahedron: <b> o&#8212;4&#8212;o&#8212;3&#8212;• </b></h3>
<p>Then finally we get the <a href="http://en.wikipedia.org/wiki/Octahedron"><b>octahedron</b></a>, with triangles as faces:</p>
<div align="center"><a href="http://en.wikipedia.org/wiki/Octahedron"><br />
<img width="300" src="https://i2.wp.com/math.ucr.edu/home/baez/mathematical/400px-octahedron.png" /></a></div>
<h3> Partial flags </h3>
<p>Can you see what&#8217;s going on with the diagrams here?</p>
<div align="center">
<table border="2">
<tr>
<td> cube
</td>
<td> <img width="140" src="https://i2.wp.com/math.ucr.edu/home/baez/mathematical/400px-cube.png" />
</td>
<td> <b> •&#8212;4&#8212;o&#8212;3&#8212;o </b>
</td>
</tr>
<tr>
<td> truncated cube
</td>
<td>  <img width="140" src="https://i2.wp.com/math.ucr.edu/home/baez/mathematical/400px-truncated_cube.png" />  </td>
<td> <b> •&#8212;4&#8212;•&#8212;3&#8212;o</b>
</td>
</tr>
<tr>
<td> cuboctahedron
</td>
<td>  <img width="140" src="https://i2.wp.com/math.ucr.edu/home/baez/mathematical/400px-cuboctahedron.png" />  </td>
<td> <b> o&#8212;4&#8212;•&#8212;3&#8212;o</b>
</td>
</tr>
<tr>
<td> truncated octahedron
</td>
<td>  <img width="140" src="https://i1.wp.com/math.ucr.edu/home/baez/mathematical/400px-truncated_octahedron.png" />  </td>
<td> <b> o&#8212;4&#8212;•&#8212;3&#8212;•</b>
</td>
</tr>
<tr>
<td> octahedron
</td>
<td>  <img width="140" src="https://i2.wp.com/math.ucr.edu/home/baez/mathematical/400px-octahedron.png" />  </td>
<td> <b> o&#8212;4&#8212;o&#8212;3&#8212;•</b>
</td>
</tr>
</table>
</div>
<p>Clearly the black dots tend to move from left to right as we move down the chart, but there&#8217;s something much cooler and more precise going on.  The black dots secretly say where the <i>corners</i> of the shapes are!</p>
<p>Let&#8217;s see how quickly I can explain this, and how quickly you can get what I&#8217;m talking about.  Remember how I defined a &#8216;flag&#8217; in <a href="https://johncarlosbaez.wordpress.com/2012/07/22/symmetry-and-the-fourth-dimension-part-3/">Part 3</a>?  </p>
<p>No?  Good, because today I&#8217;m going to call that a &#8216;complete flag&#8217;.  So, given a Platonic solid, we&#8217;ll say a <b><a href="http://en.wikipedia.org/wiki/Flag_%28geometry%29">complete flag</a></b> is a vertex, edge and face where the vertex lies on the edge and the edge lies on the face.   </p>
<p>For example, here is a complete flag for a cube:</p>
<div align="center">
<img width="300" src="https://i1.wp.com/math.ucr.edu/home/baez/mathematical/cube_with_complete_flag.png" /></div>
<p>It&#8217;s the black vertex lying on the blue edge lying on the yellow face. </p>
<p>But the term &#8216;complete flag&#8217; hints that there are also &#8216;partial flags&#8217;.  And there are!   A <b>vertex-edge flag</b> is a vertex and edge where the vertex lies on the edge.   Here&#8217;s a vertex-edge flag for the cube:</p>
<div align="center">
<img width="300" src="https://i0.wp.com/math.ucr.edu/home/baez/mathematical/cube_with_vertex_edge_flag.png" /></div>
<p>Similarly, an <b>edge-face flag</b> is an edge and a face where the edge lies on the face.  Here&#8217;s an edge-face flag for the cube:</p>
<div align="center">
<img width="300" src="https://i0.wp.com/math.ucr.edu/home/baez/mathematical/cube_with_edge_face_flag.png" /></div>
<p>You can see why they&#8217;re called partial flags: they&#8217;re different parts of a complete flag.</p>
<p>Now, fix the Coxeter diagram for the cube firmly in mind:</p>
<div align="center">
<b> V&#8212;4&#8212;E&#8212;3&#8212;F </b>
</div>
<p>V for vertex, E for edge and F for face.</p>
<p>Then:</p>
<p>&bull; a <i>cube</i> obviously has one corner for each <i>vertex</i> of the cube, so we draw it like this: </p>
<div align="center"><b> •&#8212;4&#8212;o&#8212;3&#8212;o </b></div>
<p>&bull;  a <i>truncated cube</i> has one corner for each <i>vertex-edge</i> flag of the cube, so we draw it like this: </p>
<div align="center"><b> •&#8212;4&#8212;•&#8212;3&#8212;o </b></div>
<p>&bull; a <i>cuboctahedron</i> has one corner for each <i>edge</i> of the cube, so we draw it like this: </p>
<div align="center"><b> o&#8212;4&#8212;•&#8212;3&#8212;o </b></div>
<p>&bull; a <i>truncated octahedron</i> has one corner for each <i>edge-face</i> flag of the cube, so we draw it like this: </p>
<div align="center"><b> o&#8212;4&#8212;•&#8212;3&#8212;• </b></div>
<p>&bull; an <i>octahedron</i> has one corner for each <i>face</i> of the cube, so we draw it like this:  </p>
<div align="center"><b> o&#8212;4&#8212;o&#8212;3&#8212;• </b></div>
<p>Alas, I don&#8217;t have the patience to draw all the pictures needed to explain this clearly; I&#8217;ll just grab the pictures I can get for free on Wikicommons.   Here&#8217;s how an octahedron has a corner for each face of the cube:</p>
<div align="center"><img src="https://i0.wp.com/math.ucr.edu/home/baez/mathematical/octahedron_in_cube.png" alt="" /> </div>
<p>And here&#8217;s how the cuboctahedron has a corner for each edge of the cube:</p>
<div align="center"><img width="250" src="https://i0.wp.com/math.ucr.edu/home/baez/mathematical/cuboctahedron_in_cube.png" alt="" /> </div>
<p>But these are the least interesting cases!  It&#8217;s more interesting to see how the truncated cube has one corner for each vertex-edge flag of the cube.   Do you see how it works?   You have to imagine this truncated cube sitting inside a cube:</p>
<div align="center"><a href="http://en.wikipedia.org/wiki/Truncated_cube"><br />
<img width="300" src="https://i1.wp.com/math.ucr.edu/home/baez/mathematical/400px-truncated_cube_green.jpg" /></a></div>
<p>Then, notice that the truncated cube has 2 corners on each edge of the cube, one near each end.  So, it has one corner for each vertex-edge flag of the cube!</p>
<p>Similarly, the truncated octahedron has one corner for each <i>edge-face</i> flag of the cube.  But since I don&#8217;t have a great picture to help you see that, lets use duality to change our point of view.   A face of the cube corresponds to a vertex of the octahedron.   So, think of this truncated octahedron as sitting inside an <i>octahedron</i>:</p>
<div align="center"><a href="http://en.wikipedia.org/wiki/Truncated_octahedron"><br />
<img width="300" src="https://i1.wp.com/math.ucr.edu/home/baez/mathematical/400px-truncated_octahedron_green.jpg" /></a></div>
<p>The truncated octahedron has 4 corners near each vertex of the octahedron, one on each edge touching that vertex.  In short, it has one corner for each vertex-edge flag of the octahedron.  So it&#8217;s got one corner for each edge-face flag of the cube!</p>
<p>This change of viewpoint can be justified more thoroughly:</p>
<p><b>Puzzle.</b>  The diagrams we&#8217;ve been using were based on the Coxeter diagram for the cube.  What would they look like if we based them on the Coxeter diagram for the octahedron instead?</p>
<p>By the way, the pretty pictures of solids with brass balls at the vertices were made by Tom Ruen using Robert Webb&#8217;s <a href="http://www.software3d.com/Stella.php">Stella</a> software.  They&#8217;re available on Wikicommons, and you can find most of them by clicking on the images here and looking around on the Wikipedia articles you&#8217;ll reach that way.  On this blog, I try hard to make most images take you to more information when you click on them.</p>
]]></html><thumbnail_url><![CDATA[https://i0.wp.com/math.ucr.edu/home/baez/mathematical/tetrahedron_spinning.gif?fit=440%2C330]]></thumbnail_url><thumbnail_height><![CDATA[256]]></thumbnail_height><thumbnail_width><![CDATA[256]]></thumbnail_width></oembed>