FRIEZE SYMMETRY (Part I) ------------------------ In architecture, the term frieze refers to an ornately decorated band running along the the top of a wall or other structure. When the decoration has a repetitive nature, that raise the mathematical question of classifying its symmetry type. It turns out that a frieze may possess one of exactly 7 different symmetry types. One may relate frieze symmetry to wallpaper symmetry in several ways: 1) One may view frieze symmetry as a ``toy'' version of wallpaper symmetry, a somewhat simpler context where we encounter similar issues. Thus we can study friezes to reinforce our understanding of the plane symmetry classification; 2) One may extend a frieze pattern to a plane pattern in two distinct ways (see below), so one may use the classification of plane symmetry types to derive the classification for frieze symmetry types; 3) One may relate frieze symmetry types to symmetric patterns on the sphere. THE SEVEN FRIEZE TYPES AND A CATCHY WAY TO REMEMBER THEM -------------------------------------------------------- A trail of footprints counts as a frieze decoration. The symmetry type depends on the nature of the stride, as follows: HOP ----------------------------------------------------------------------------------------------- ..... o ..... o ..... o ..... o ..... o . . o . . o . . o . . o . . o ... . o ... . o ... . o ... . o ... . o .. o .. o .. o .. o .. o ----------------------------------------------------------------------------------------------- JUMP ------------------------------------------------------------------------------------------- ..... o ..... o ..... o ..... o ..... o ..... o ..... o ..... o ..... o . . o . . o . . o . . o . . o . . o . . o . . o . . o ... . o ... . o ... . o ... . o ... . o ... . o ... . o ... . o ... . o .. o .. o .. o .. o .. o .. o .. o .. o .. o .. o .. o .. o .. o .. o .. o .. o .. o .. o ... . o ... . o ... . o ... . o ... . o ... . o ... . o ... . o ... . o . . o . . o . . o . . o . . o . . o . . o . . o . . o ..... o ..... o ..... o ..... o ..... o ..... o ..... o ..... o ..... o ------------------------------------------------------------------------------------------- SIDE-STEP or SIDLE ------------------------------------------------------------------------------------------- oooo oooo oooo oooo oooo oooo oooo oooo oooo oooo oooo oooo oooo oooo .. .. .. .. .. .. .. .. .. .. .. .. .. .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ------------------------------------------------------------------------------------------- WALK ----------------------------------------------------------------------------------------------- ..... o ..... o ..... o ..... o ..... o . . o . . o . . o . . o . . o ... . o ... . o ... . o ... . o ... . o .. o .. o .. o .. o .. o .. o .. o .. o .. o ... . o ... . o ... . o ... . o . . o . . o . . o . . o ..... o ..... o ..... o ..... o ----------------------------------------------------------------------------------------------- SPIN-JUMP ----------------------------------------------------------------------------------- ..... o o ..... ..... o o ..... ..... o o ..... ..... o . . o o . . . . o o . . . . o o . . . . o ... . o o . ... ... . o o . ... ... . o o . ... ... . o .. o o .. .. o o .. .. o o .. .. o .. o o .. .. o o .. .. o o .. .. o ... . o o . ... ... . o o . ... ... . o o . ... ... . o . . o o . . . . o o . . . . o o . . . . o ..... o o ..... ..... o o ..... ..... o o ..... ..... o ----------------------------------------------------------------------------------- SPIN-HOP ----------------------------------------------------------------------------------- ..... o ..... o ..... o ..... o . . o . . o . . o . . o ... . o ... . o ... . o ... . o .. o .. o .. o .. o o .. o .. o .. o . ... o . ... o . ... o . . o . . o . . o ..... o ..... o ..... ----------------------------------------------------------------------------------- SPIN-SIDLE ---------------------------------------------------------------------------------------- oooo oooo . . oooo oooo . . oooo oooo . . oooo .. .. . . . . .. .. . . . . .. .. . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. .. . . . . .. .. . . . . .. .. . . . . oooo oooo . . oooo oooo . . oooo oooo . ---------------------------------------------------------------------------------------- At this point, you should hardly find it obvious that no other symmetry type occurs! Exercise: Determine the symmetry type (one of the above)of the following footprint frieze: ----------------------------------------------------------------------------------- .. o o .. .. o o .. .. o o .. .. o ... . o o . ... ... . o o . ... ... . o o . ... ... . o . . o o . . . . o o . . . . o o . . . . o ..... o o ..... ..... o o ..... ..... o o ..... ..... o ----------------------------------------------------------------------------------- (Beside helping us to remember the 7 types, this nomenclature suggests that frieze symmetry can model the symmetries of dance steps, or equally well, repetitive musical motifs.) Now that we've met the 7 types, let's review some of their features. HOP possesses only translation symmetry to the left and right. JUMP possesses a line of reflection running along the middle of the strip. SIDLE possesses two sorts of reflection lines running perpendicular to the direction of the strip. Reflection occur between consecutive footprint. One sort lies close to the big toes, the other close to the little toes. WALK possesses a line of glide reflection running along the middle of the strip. SPIN-JUMP possesses a line of reflection running along the middle of the strip. as well as two types of perpendicular reflection lines, some near toes, some near heels. SPIN-HOP has no reflections or glide reflections, but does have two types of order 2 gyration points. Some lie midway between nearby big toes, some midway between nearby heels. SPIN-SIDLE has perpendicular reflection lines running midway between consecutive footprints that point the same way, and gyration points of order 2 that lie midway between oppositely oriented consecutive footprints. (A line of glide reflection does run down the middle of the strip, but since that line contain gyration points, it will not play any further role in our analysis). NEW NAMES --------- Just as will plane symmetric patterns, extracting a template and folding it into a polyhedron form a crucial step in our understand of frieze symmetry. For reason that shall only become clear as we go on, we now choose to think of each strip as sitting inside a plane and work with the entire plane. Thus the templates will have infinite area; this marks frieze symmetry as distinct from the plane symmetries we've already studied. In the case of our standard examples, each template should contain one footprint. The following diagram shows a template for HOP: | | | Typical template | | | | | ..... o | ..... o | ..... o ..... o . . o | . . o | . . o . . o ... . o | ... . o | ... . o ... . o .. o | .. o | .. o .. o | | | (extends up and | | down infinitely) | | | | | To form a(n infinite) polyhedron from a template we just tape together the edges to form an infinite cylinder. Let us regard the two ends of the cylinder as ``cone points'' of zero angle and infinite order. For now you should just take this as a suggestive convention. (Of course there exists no point located off at infinity at the ends of the cylinder, so we had to put _cone points_ in quotes.) Since a typewriter can't make the standard eight-on-its-side symbol for infinite, we'll abbreviate it as Inf. So with these conventions we get a new name for HOP imitating our plane symmetry notation: Inf Inf Now we turn to JUMP: | | | Template| | | | | | | | | | | ..... o ..... o| ..... o| ..... o ..... o ..... o ..... o ..... o ..... o . . o . . o|. . o|. . o . . o . . o . . o . . o . . o ... . o ... . o| ... . o| ... . o ... . o ... . o ... . o ... . o ... . o .. o .. o| .. o| .. o .. o .. o .. o .. o .. o ----------- .. o .. o .. o .. o .. o .. o .. o .. o .. o ... . o ... . o ... . o ... . o ... . o ... . o ... . o ... . o ... . o . . o . . o . . o . . o . . o . . o . . o . . o . . o ..... o ..... o ..... o ..... o ..... o ..... o ..... o ..... o ..... o This time we get a cylinder infinite in one direction and terminating at a reflective edge in the other. This suggests the notation Inf * As for SIDLE, the template we get consists of an infinite strip with reflective edges on each side. Now we should regard the ends as ``corner points'' or zero angle and infinite order. | | | | | | oooo oooo | oooo|oooo oooo oooo oooo oooo oooo oooo oooo oooo oooo oooo .. .. | .. | .. .. .. .. .. .. .. .. .. .. .. . . . . | . .|. . . . . . . . . . . . . . . . . . . . . . . . . . | . .|. . . . . . . . . . . . . . . . . . . . . . . . . . | . . | . . . . . . . . . . . . . . . . . . . . . . . . . . | . . | . . . . . . . . . . . . . . . . . . . . . . . . . . | . . | . . . . . . . . . . . . . . . . . . . . . . . . | . | . . . . . . . . . . . | | | | | | This suggests the notation * Inf Inf Why just one * when the pattern has reflection lines of two types? In the context of plane symmetric patterns with finite templates, parallel lines never met, but here we regard them as ``meeting at infinity'' with ``angle zero''. So the two line form parts of the same infinite boundary. If this seems like cheating, or seems arbitrary, have patience, the sense behind it all will emerge gradually. Now a template for WALK reminds us of the template for HOP | | | | | | ..... o | ..... o| ..... o ..... o ..... o . . o | . . o| . . o . . o . . o ... . o | ... . o| ... . o ... . o ... . o .. o | .. o| .. o .. o .. o .. o| | .. o .. o .. o ... . o| | ... . o ... . o ... . o . . o| |. . o . . o . . o ..... o| | ..... o ..... o ..... o | | | | | | except now we must tape together with a twist. The polyhedron we get clearly contains a Mobius strip. It also has an infinite order ``cone point'', but we won't try to visualize it. In any case we should give it the name Inf x SPIN-JUMP presents no difficulties. The template looks like | | | | | | ..... o | o ..... | ..... o o ..... ..... o o ..... ..... o . . o | o . . | . . o o . . . . o o . . . . o ... . o | o . ... | ... . o o . ... ... . o o . ... ... . o .. o | o .. | .. o o .. .. o o .. .. o ------------ .. o o .. .. o o .. .. o o .. .. o ... . o o . ... ... . o o . ... ... . o o . ... ... . o . . o o . . . . o o . . . . o o . . . . o ..... o o ..... ..... o o ..... ..... o o ..... ..... o with all edges reflective. We have, plainly visible, two order 2 corner points, and a zero angle infinite order ``corner point'' at the end of the template. Thus the name * 2 2 Inf With SPIN-HOP a template might look or like this. either like this | | | | | | | | | | | | | | | | | | | | ..... o | ..... o | ..... o | | ..... o . . o | . . o | . . o | | . . o ... . o | ... . o | ... . o | | ... . o .. o | .. o | .. o | | .. o ------------------------- | | o .. o .. | o .. | o . ... o . ... | o . ... | o . . o . . | o . . | o ..... o ..... | o ..... | | | | | To form the polyhedron we would either crease the first template vertically and tape the edges together, or crease the second template horizontally and tape. Either way we get an infinite (semi-)rectangular pillowcase (this illustrates why we emphasize polyhedra over templates) and the name 2 2 Inf Something similar happens with SPIN-SIDLE. We could have a template | | | | | | | | | | oooo oooo . . oooo | oooo | . . oooo oooo . . .. .. . . . . .. | .. | . . . . .. .. . . . . . . . . . . . . . . | . . | . . . . . . . . . . . . . . . . . . . . . . | . . | . . . . . . . . . . . . . . . . . . . . . . | . . | . . . . . . . . . . . . . . . . . . . . . . | . . | . . . . . . . . . . . . . . . . .. .. . . | . . | .. .. . . . . .. .. . . oooo oooo . | . | oooo oooo . . oooo oooo | | | | | | | | | | and again we should crease horizontally. But now we have a reflective edge on the left which we shouldn't tape, so we tape only on the right. We'll get an order 2 cone point on the right and an infinite order ``corner point'' at the end of the strip. Hence the name 2 * Inf THE PARALLELS WITH PLANE SYMMETRY ---------------------------------_ A look at the names we've just derived shows that the satisfy a logic familiar to us for plane symmetry. We have two names which feature only gyration points: 2 2 Inf Inf Inf We can place a * in front of each name * 2 2 Inf * Inf Inf and then move pairs of corner points left, across the * , to make a gyration point. Inf * 2 * Inf Finally we can replace a * with no corner points by a x : Inf x The way this story parallels our previous one suggests that we're doing something right. In fact we can push the parallel further. Returning to the idea of building polyhedra by modifying a sphere and ``charging'' ourselves for various features, we could charge $1 for a ``cone point'' Inf and $.50 for a ``corner point'' Inf and we'd get exactly these 7 ways to spend $2 and use one of the new coins. Since a cone point Inf costs the same as a * , we can reuse our old analysis. We had three types with at least one * with no corner points: **, *x, and 22*. Changing one or more *'s to Inf's gives Inf Inf , Inf * , Inf x and 2 2 Inf. Now we can enumerate the posibilities that use corner point Inf from these, just as before. So what's the status of all this?? We still haven't proved the completeness of our classification. We have no Euler theorem or Gauss-Bonnet theorem for infinite polyhedra, so we can't simply lift the old argument. Still, everything appears to work the same way, and that deserves an explanation.