Sedenion Geometry
As seen elsewhere, the imaginary elements and triads of the [twisted] octonions form a projective geometry with points and lines, namely, the Fano plane. Likewise, the imaginary units and quaternionic groupings (triads) of the various types of sedenions can be considered as points and lines, and heptads have some analogies to planes.
Consider first the construction of triad-based multiplications tables for systems with 7 imaginaries, such that any pair of units is in a unique triad. An exhaustive analysis shows that all possible groupings of units into such triads have the same basic form:
- each element is in 3 triads
- there are 7 distinct triads
- any 2 distinct triads intersect in a single unit
- all such groupings are isomorphic, under permutation of the indices labelling the units
The proofs:
- each element is in 6 pairs of elements, and each triad containing the element "uses up" 2 of those pairs (since each pair is in a unique triad), yielding 3 triads per unit
- 3 triads per element, × 7 elements = 21, but each triad is hereby counted thrice, once for each element it contains, so there are 3×7/3 = 7 distinct triads
- for triads t1, t2 let N=o(t1∩t2), let e∈(t1-t2), and let T={ef:f∈(t2-t1)}; then T∩t1 = T∩t2 = ∅ and o(T) = o(t1-t2) =o (t2-t1) = 3-N, so
o(T∪(t1-t2)∪(t2-t1)) = 3(3-N) ≤ 7 ⇒ N ≥ 1
Now consider hypercomplex systems having fifteen imaginary units grouped into overlapping triads, such that each pair of units determines a unique triad. The resulting multiplication might or might not have closed subalgebras spanned by 7 imaginaries ("heptads"); when heptads are present, there are 1, 3, 7, or 15 of them (for reasons reminiscent of the XOR or Cayley-Dickson doubling construction, as will be seen). Some basic triad properties are independent of the presence or absence of heptads, and will be investigated first. In the following, a heptad h = {ei∈I : o(I) = 7} will loosely be described as "closed under multiplication" if
ei, ej ∈ h ⇒ eiej ∈ h∪(-h)∪{±e0}
where -h = {-ei : i∈I} and e0 is the multiplicative identity.
Properties for Any Number of Heptads
Combinatoric arguments easily establish certain basic properties of all quaternionic groupings of 15 imaginary elements, e.g.:
- each imaginary unit is in 7 triads
- there are a total of 35 quaternionic triads
- each triad intersects 18 others, is disjoint with 16 others
- if there exist more than one heptad, 2 distinct heptads intersect in a triad
- any triad can be in at most 3 heptads
- each triad intersects any heptad in 1 or 3 basis elements
The proofs are:
- each imaginary ei is in 14 pairs; each pair is in exactly one triad; each triad containing ei contains 2 of these pairs; hence, there are exactly 7 triads containing ei
- each of the 15 elements ei occurs in 7 triads, giving 105 = 7×15; each triad is thereby counted 3 times, once for each ei it contains, thus giving 105/3 = 35 distinct triads
- triad (ei,ej,ek) intersects all other triads containing ei, ej, or ek, and only those; there are 6 such triads for each element, all distinct, hence 18 in total; thus there are 35 - 18 - 1 other triads not containing any of ei, ej, or ek, i.e. 16 triads disjoint with the given triad
- let heptads h1, h2 have intersection h1∩h2 containing N elements; let Hi = hi - (h1∩h2), so o(Hi) = 7-N; for E∈H1, EH2 is a set of 7-N distinct elements ∉h1 and ∉h2; thus o(h1∪h2∪EH2) = N + 3(7-N) = 21-2N ≤ 15 distinct elements, so that N ≥ 3. Moreover, if ei, ej∈h1∩h2, then eiej∈h1∩h2, so if h1∩h2 comprises more than a single triad, it must also contain all products of the contained elements, or 7 elements in all, so h1∩h2 = h1 = h2.
- if a triad t is in N heptads, each heptad has 4 elements not in any other of the heptads containing t, for a total of 4N + 3 ≤ 15 distinct units
- let N = o(t∩h) for triad t⊄heptad h; let e∈t-h, and H = {ef: f∈(h-t)}; then o(H) = 7-N, and t∩H = h∩H = ∅; then 15 ≥ o(t)+o(h-t)+o(H) = 3+(7-N)+(7-N) ⇒ N≥1; since h is closed under multiplication, N>1 ⇒ N=3
Systems with 0 Heptads
The choice of triads:
0: e1,e2,e3 | 1: e1,e4,e5 | 2: e1,e6,e7 | 3: e1,e8,e9 | 4: e1,e10,e11 |
5: e1,e12,e13 | 6: e1,e14,e15 | 7: e2,e4,e6 | 8: e2,e5,e7 | 9: e2,e8,e10 |
10: e2,e9,e11 | 11: e2,e12,e14 | 12: e2,e13,e15 | 13: e3,e4,e7 | 14: e3,e5,e10 |
15: e3,e6,e15 | 16: e3,e8,e14 | 17: e3,e9,e13 | 18: e3,e11,e12 | 19: e4,e8,e15 |
20: e4,e9,e12 | 21: e4,e10,e13 | 22: e4,e11,e14 | 23: e5,e6,e12 | 24: e5,e8,e13 |
25: e5,e9,e14 | 26: e5,e11,e15 | 27: e6,e8,e11 | 28: e6,e9,e10 | 29: e6,e13,e14 |
30: e7,e8,e12 | 31: e7,e9,e15 | 32: e7,e10,e14 | 33: e7,e11,e13 | 34: e10,e12,e15 |
produces (when all triads are made right-handed) the multiplication table below.
| e0 | e1 | e2 | e3 | e4 | e5 | e6 | e7 | e8 | e9 | e10 | e11 | e12 | e13 | e14 | e15 |
e0 | e0 | e1 | e2 | e3 | e4 | e5 | e6 | e7 | e8 | e9 | e10 | e11 | e12 | e13 | e14 | e15 |
e1 | e1 | -e0 | e3 | -e2 | e5 | -e4 | e7 | -e6 | e9 | -e8 | e11 | -e10 | e13 | -e12 | e15 | -e14 |
e2 | e2 | -e3 | -e0 | e1 | e6 | e7 | -e4 | -e5 | e10 | e11 | -e8 | -e9 | e14 | e15 | -e12 | -e13 |
e3 | e3 | e2 | -e1 | -e0 | e7 | e10 | e15 | -e4 | e14 | e13 | -e5 | e12 | -e11 | -e9 | -e8 | -e6 |
e4 | e4 | -e5 | -e6 | -e7 | -e0 | e1 | e2 | e3 | e15 | e12 | e13 | e14 | -e9 | -e10 | -e11 | -e8 |
e5 | e5 | e4 | -e7 | -e10 | -e1 | -e0 | e12 | e2 | e13 | e14 | e3 | e15 | -e6 | -e8 | -e9 | -e11 |
e6 | e6 | -e7 | e4 | -e15 | -e2 | -e12 | -e0 | e1 | e11 | e10 | -e9 | -e8 | e5 | e14 | -e13 | e3 |
e7 | e7 | e6 | e5 | e4 | -e3 | -e2 | -e1 | -e0 | e12 | e15 | e14 | e13 | -e8 | -e11 | -e10 | -e9 |
e8 | e8 | -e9 | -e10 | -e14 | -e15 | -e13 | -e11 | -e12 | -e0 | e1 | e2 | e6 | e7 | e5 | e3 | e4 |
e9 | e9 | e8 | -e11 | -e13 | -e12 | -e14 | -e10 | -e15 | -e1 | -e0 | e6 | e2 | e4 | e3 | e5 | e7 |
e10 | e10 | -e11 | e8 | e5 | -e13 | -e3 | e9 | -e14 | -e2 | -e6 | -e0 | e1 | e15 | e4 | e7 | -e12 |
e11 | e11 | e10 | e9 | -e12 | -e14 | -e15 | e8 | -e13 | -e6 | -e2 | -e1 | -e0 | e3 | e7 | e4 | e5 |
e12 | e12 | -e13 | -e14 | e11 | e9 | e6 | -e5 | e8 | -e7 | -e4 | -e15 | -e3 | -e0 | e1 | e2 | e10 |
e13 | e13 | e12 | -e15 | e9 | e10 | e8 | -e14 | e11 | -e5 | -e3 | -e4 | -e7 | -e1 | -e0 | e6 | e2 |
e14 | e14 | -e15 | e12 | e8 | e11 | e9 | e13 | e10 | -e3 | -e5 | -e7 | -e4 | -e2 | -e6 | -e0 | e1 |
e15 | e15 | e14 | e13 | e6 | e8 | e11 | -e3 | e9 | -e4 | -e7 | e12 | -e5 | -e10 | -e2 | -e1 | -e0 |
It is easy to verify that any 3 elements not contained in a single triad will multiplicatively generate the full algebra; e.g. starting with e1, e2, e4 gives e1e2=e3, e1e4=e5, e2e4=e6; these in turn give e3e5=e10, e5e6=e12, e3e6=e15; continuing eventually generates each of the 15 units. The C program sedhepts.c will read in a specified list of triads (one triad per line after a "signmask" in the first line) and display the heptadic structure of the resulting algebra.
Testing the 15! permutations of the 15 elements reveals that only the identity permutation leaves this triad set unchanged; hence there are 15! distinct multiplication tables obtained from this one via permutation of the indices of the units.
Systems with 1 Heptad
The choice of triads
0: e1,e2,e3 | 1: e1,e4,e5 | 2: e1,e6,e7 | 3: e1,e8,e9 | 4: e1,e10,e11 |
5: e1,e12,e13 | 6: e1,e14,e15 | 7: e2,e4,e6 | 8: e2,e5,e7 | 9: e2,e8,e10 |
10: e2,e9,e11 | 11: e2,e12,e14 | 12: e2,e13,e15 | 13: e3,e4,e7 | 14: e3,e5,e6 |
15: e3,e8,e11 | 16: e3,e9,e12 | 17: e3,e10,e15 | 18: e3,e13,e14 | 19: e4,e8,e12 |
20: e4,e9,e13 | 21: e4,e10,e14 | 22: e4,e11,e15 | 23: e5,e8,e14 | 24: e5,e9,e10 |
25: e5,e11,e13 | 26: e5,e12,e15 | 27: e6,e8,e15 | 28: e6,e9,e14 | 29: e6,e10,e13 |
30: e6,e11,e12 | 31: e7,e8,e13 | 32: e7,e9,e15 | 33: e7,e10,e12 | 34: e7,e11,e14 |
yields (when all triads are right-handed) an algebra with exactly one multiplicatively closed heptad subalgebra, containing units {e1, e2, e3, e4, e5, e6, e7}. The multiplication table (with signs from 35-bit signmask=0) is:
| e0 | e1 | e2 | e3 | e4 | e5 | e6 | e7 | e8 | e9 | e10 | e11 | e12 | e13 | e14 | e15 |
e0 | e0 | e1 | e2 | e3 | e4 | e5 | e6 | e7 | e8 | e9 | e10 | e11 | e12 | e13 | e14 | e15 |
e1 | e1 | -e0 | e3 | -e2 | e5 | -e4 | e7 | -e6 | e9 | -e8 | e11 | -e10 | e13 | -e12 | e15 | -e14 |
e2 | e2 | -e3 | -e0 | e1 | e6 | e7 | -e4 | -e5 | e10 | e11 | -e8 | -e9 | e14 | e15 | -e12 | -e13 |
e3 | e3 | e2 | -e1 | -e0 | e7 | e6 | -e5 | -e4 | e11 | e12 | e15 | -e8 | -e9 | e14 | -e13 | -e10 |
e4 | e4 | -e5 | -e6 | -e7 | -e0 | e1 | e2 | e3 | e12 | e13 | e14 | e15 | -e8 | -e9 | -e10 | -e11 |
e5 | e5 | e4 | -e7 | -e6 | -e1 | -e0 | e3 | e2 | e14 | e10 | -e9 | e13 | e15 | -e11 | -e8 | -e12 |
e6 | e6 | -e7 | e4 | e5 | -e2 | -e3 | -e0 | e1 | e15 | e14 | e13 | e12 | -e11 | -e10 | -e9 | -e8 |
e7 | e7 | e6 | e5 | e4 | -e3 | -e2 | -e1 | -e0 | e13 | e15 | e12 | e14 | -e10 | -e8 | -e11 | -e9 |
e8 | e8 | -e9 | -e10 | -e11 | -e12 | -e14 | -e15 | -e13 | -e0 | e1 | e2 | e3 | e4 | e7 | e5 | e6 |
e9 | e9 | e8 | -e11 | -e12 | -e13 | -e10 | -e14 | -e15 | -e1 | -e0 | e5 | e2 | e3 | e4 | e6 | e7 |
e10 | e10 | -e11 | e8 | -e15 | -e14 | e9 | -e13 | -e12 | -e2 | -e5 | -e0 | e1 | e7 | e6 | e4 | e3 |
e11 | e11 | e10 | e9 | e8 | -e15 | -e13 | -e12 | -e14 | -e3 | -e2 | -e1 | -e0 | e6 | e5 | e7 | e4 |
e12 | e12 | -e13 | -e14 | e9 | e8 | -e15 | e11 | e10 | -e4 | -e3 | -e7 | -e6 | -e0 | e1 | e2 | e5 |
e13 | e13 | e12 | -e15 | -e14 | e9 | e11 | e10 | e8 | -e7 | -e4 | -e6 | -e5 | -e1 | -e0 | e3 | e2 |
e14 | e14 | -e15 | e12 | e13 | e10 | e8 | e9 | e11 | -e5 | -e6 | -e4 | -e7 | -e2 | -e3 | -e0 | e1 |
e15 | e15 | e14 | e13 | e10 | e11 | e12 | e8 | e9 | -e6 | -e7 | -e3 | -e4 | -e5 | -e2 | -e1 | -e0 |
Testing the 15! permutations of unit indices yields an 8-element group that leaves the multiplication table invariant, namely, the permutations:
#1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 I
#16704 1 2 3 4 5 6 7 11 10 9 8 15 14 13 12 a
#28784 1 2 3 4 5 6 7 13 12 15 14 9 8 11 10 b
#35153 1 2 3 4 5 6 7 14 15 12 13 10 11 8 9 c
#26911186439 1 6 7 4 5 2 3 9 8 14 15 13 12 10 11 d
#26911192746 1 6 7 4 5 2 3 10 11 13 12 14 15 9 8 e
#26911204250 1 6 7 4 5 2 3 12 13 11 10 8 9 15 14 f
#26911220407 1 6 7 4 5 2 3 15 14 8 9 11 10 12 13 g
where I is the identity, and using the 1-letter labels, the group multiplication table is:
| a | b | c | d | e | f | g |
a | I | c | b | g | f | e | d |
b | c | I | a | f | g | d | e |
c | b | a | I | e | d | g | f |
d | e | f | g | I | a | b | c |
e | d | g | f | c | b | a | I |
f | g | d | e | b | c | I | a |
g | f | e | d | a | I | c | b |
Systems with 2 Heptads
A system S with 15 imaginary units and at least 2 distinct heptads h1 and h2 must necessarily have at least 3 distinct heptads. By the lemma above, there is a triad T = h1∩h2. Each ei∈T is also in 2 other triads ⊂ h1, and in 2 further distinct triads ⊂ h2, for a total of 5 (counting T). Since ei is in a total of 7 triads, there are 2 further triads ∋ ei, each containing 2 of the 4 elements of S - (h1∪h2). It easily follows that T∪(S - (h1∪h2)) is also a heptad closed under multiplication.
Systems with 3 Heptads
The choice of triads:
0: e1,e2,e3 | 1: e1,e4,e5 | 2: e1,e6,e7 | 3: e1,e8,e9 | 4: e1,e10,e11 |
5: e1,e12,e13 | 6: e1,e14,e15 | 7: e2,e4,e6 | 8: e2,e5,e7 | 9: e2,e8,e10 |
10: e2,e9,e11 | 11: e2,e12,e14 | 12: e2,e13,e15 | 13: e3,e4,e7 | 14: e3,e5,e6 |
15: e3,e8,e11 | 16: e3,e9,e10 | 17: e3,e12,e15 | 18: e3,e13,e14 | 19: e4,e8,e12 |
20: e4,e9,e13 | 21: e4,e10,e14 | 22: e4,e11,e15 | 23: e5,e8,e13 | 24: e5,e9,e14 |
25: e5,e10,e15 | 26: e5,e11,e12 | 27: e6,e8,e15 | 28: e6,e9,e12 | 29: e6,e10,e13 |
30: e6,e11,e14 | 31: e7,e8,e14 | 32: e7,e9,e15 | 33: e7,e10,e12 | 34: e7,e11,e13 |
produces a system with 3 distinct subalgebra heptads, namely, the heptads composed of elements:
- e1, e2, e3, e4, e5, e6, e7,
- e1, e2, e3, e8, e9, e10, e11, and
- e1, e2, e3, e12, e13, e14, e15.
Note the presence of a common triad (i.e. e1, e2, e3), as expected from the discussion above for Systems with 2 Heptads. The multiplication table is (for all triads right-handed):
| e0 | e1 | e2 | e3 | e4 | e5 | e6 | e7 | e8 | e9 | e10 | e11 | e12 | e13 | e14 | e15 |
e0 | e0 | e1 | e2 | e3 | e4 | e5 | e6 | e7 | e8 | e9 | e10 | e11 | e12 | e13 | e14 | e15 |
e1 | e1 | -e0 | e3 | -e2 | e5 | -e4 | e7 | -e6 | e9 | -e8 | e11 | -e10 | e13 | -e12 | e15 | -e14 |
e2 | e2 | -e3 | -e0 | e1 | e6 | e7 | -e4 | -e5 | e10 | e11 | -e8 | -e9 | e14 | e15 | -e12 | -e13 |
e3 | e3 | e2 | -e1 | -e0 | e7 | e6 | -e5 | -e4 | e11 | e10 | -e9 | -e8 | e15 | e14 | -e13 | -e12 |
e4 | e4 | -e5 | -e6 | -e7 | -e0 | e1 | e2 | e3 | e12 | e13 | e14 | e15 | -e8 | -e9 | -e10 | -e11 |
e5 | e5 | e4 | -e7 | -e6 | -e1 | -e0 | e3 | e2 | e13 | e14 | e15 | e12 | -e11 | -e8 | -e9 | -e10 |
e6 | e6 | -e7 | e4 | e5 | -e2 | -e3 | -e0 | e1 | e15 | e12 | e13 | e14 | -e9 | -e10 | -e11 | -e8 |
e7 | e7 | e6 | e5 | e4 | -e3 | -e2 | -e1 | -e0 | e14 | e15 | e12 | e13 | -e10 | -e11 | -e8 | -e9 |
e8 | e8 | -e9 | -e10 | -e11 | -e12 | -e13 | -e15 | -e14 | -e0 | e1 | e2 | e3 | e4 | e5 | e7 | e6 |
e9 | e9 | e8 | -e11 | -e10 | -e13 | -e14 | -e12 | -e15 | -e1 | -e0 | e3 | e2 | e6 | e4 | e5 | e7 |
e10 | e10 | -e11 | e8 | e9 | -e14 | -e15 | -e13 | -e12 | -e2 | -e3 | -e0 | e1 | e7 | e6 | e4 | e5 |
e11 | e11 | e10 | e9 | e8 | -e15 | -e12 | -e14 | -e13 | -e3 | -e2 | -e1 | -e0 | e5 | e7 | e6 | e4 |
e12 | e12 | -e13 | -e14 | -e15 | e8 | e11 | e9 | e10 | -e4 | -e6 | -e7 | -e5 | -e0 | e1 | e2 | e3 |
e13 | e13 | e12 | -e15 | -e14 | e9 | e8 | e10 | e11 | -e5 | -e4 | -e6 | -e7 | -e1 | -e0 | e3 | e2 |
e14 | e14 | -e15 | e12 | e13 | e10 | e9 | e11 | e8 | -e7 | -e5 | -e4 | -e6 | -e2 | -e3 | -e0 | e1 |
e15 | e15 | e14 | e13 | e12 | e11 | e10 | e8 | e9 | -e6 | -e7 | -e5 | -e4 | -e3 | -e2 | -e1 | -e0 |
Testing the 15! permutations of unit indices yields an 8-element group that leaves the multiplication table invariant, namely, the permutations:
#1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 I
#11537 1 2 3 4 5 6 7 10 11 8 9 14 15 12 13 A
#28784 1 2 3 4 5 6 7 13 12 15 14 9 8 11 10 B
#40320 1 2 3 4 5 6 7 15 14 13 12 11 10 9 8 C
#127370897 1 2 3 7 6 5 4 8 9 10 11 14 15 12 13 D
#127382401 1 2 3 7 6 5 4 10 11 8 9 12 13 14 15 E
#127399680 1 2 3 7 6 5 4 13 12 15 14 11 10 9 8 F
#127411184 1 2 3 7 6 5 4 15 14 13 12 9 8 11 10 G
where I is the identity, and using the 1-letter labels, the group multiplication table is:
| A | B | C | D | E | F | G |
A | I | C | B | E | D | G | F |
B | C | I | A | G | F | E | D |
C | B | A | I | F | G | D | E |
D | E | F | G | I | A | B | C |
E | D | G | F | A | I | C | B |
F | G | D | E | C | B | A | I |
G | F | E | D | B | C | I | A |
Systems with 4, 5, or 6 Heptads
As a corollary to the arguments in Systems with 2 Heptads above, any system S with more than 3 distinct heptads must have at least 7. Let hi, 1≤i≤4 be 4 distinct heptads ⊂ S. Let triad tij = hi∩hj. Then either t12 ⊄ h3 or t12 ⊄ h4, so let h be the third heptad ⊂ S containing t12 (existence guaranteed by arguments in Systems with 2 Heptads above), and let H be an hi such that t12 ⊄ H. Then each of the triads H∩h1, H∩h2, H∩h are distinct, and for each there is yet another heptad which must also be ⊂ S. Note how adjoining H to h1, h2, h generates 3 more heptads, analogously to the action of adjoining a new imaginary unit to the quaternions in the Cayley-Dickson construction.
Systems with 7 Heptads
The heptads
- e1, e2, e3, e4, e5, e6, e7,
- e1, e2, e3, e8, e9, e10, e11,
- e1, e2, e3, e12, e13, e14, e15,
- e1, e4, e5, e8, e9, e12, e13,
- e1, e4, e5, e10, e11, e14, e15,
- e1, e6, e7, e8, e9, e14, e15,
- e1, e6, e7, e10, e11, e12, e13
are subalgebras for the triad choice:
0: e1,e2,e3 | 1: e1,e4,e5 | 2: e1,e6,e7 | 3: e1,e8,e9 | 4: e1,e10,e11 |
5: e1,e12,e13 | 6: e1,e14,e15 | 7: e2,e4,e6 | 8: e2,e5,e7 | 9: e2,e8,e10 |
10: e2,e9,e11 | 11: e2,e12,e14 | 12: e2,e13,e15 | 13: e3,e4,e7 | 14: e3,e5,e6 |
15: e3,e8,e11 | 16: e3,e9,e10 | 17: e3,e12,e15 | 18: e3,e13,e14 | 19: e4,e8,e12 |
20: e4,e9,e13 | 21: e4,e10,e14 | 22: e4,e11,e15 | 23: e5,e8,e13 | 24: e5,e9,e12 |
25: e5,e10,e15 | 26: e5,e11,e14 | 27: e6,e8,e14 | 28: e6,e9,e15 | 29: e6,e10,e13 |
30: e6,e11,e12 | 31: e7,e8,e15 | 32: e7,e9,e14 | 33: e7,e10,e12 | 34: e7,e11,e13 |
which produces the multiplication table (signmask = 0):
| e0 | e1 | e2 | e3 | e4 | e5 | e6 | e7 | e8 | e9 | e10 | e11 | e12 | e13 | e14 | e15 |
e0 | e0 | e1 | e2 | e3 | e4 | e5 | e6 | e7 | e8 | e9 | e10 | e11 | e12 | e13 | e14 | e15 |
e1 | e1 | -e0 | e3 | -e2 | e5 | -e4 | e7 | -e6 | e9 | -e8 | e11 | -e10 | e13 | -e12 | e15 | -e14 |
e2 | e2 | -e3 | -e0 | e1 | e6 | e7 | -e4 | -e5 | e10 | e11 | -e8 | -e9 | e14 | e15 | -e12 | -e13 |
e3 | e3 | e2 | -e1 | -e0 | e7 | e6 | -e5 | -e4 | e11 | e10 | -e9 | -e8 | e15 | e14 | -e13 | -e12 |
e4 | e4 | -e5 | -e6 | -e7 | -e0 | e1 | e2 | e3 | e12 | e13 | e14 | e15 | -e8 | -e9 | -e10 | -e11 |
e5 | e5 | e4 | -e7 | -e6 | -e1 | -e0 | e3 | e2 | e13 | e12 | e15 | e14 | -e9 | -e8 | -e11 | -e10 |
e6 | e6 | -e7 | e4 | e5 | -e2 | -e3 | -e0 | e1 | e14 | e15 | e13 | e12 | -e11 | -e10 | -e8 | -e9 |
e7 | e7 | e6 | e5 | e4 | -e3 | -e2 | -e1 | -e0 | e15 | e14 | e12 | e13 | -e10 | -e11 | -e9 | -e8 |
e8 | e8 | -e9 | -e10 | -e11 | -e12 | -e13 | -e14 | -e15 | -e0 | e1 | e2 | e3 | e4 | e5 | e6 | e7 |
e9 | e9 | e8 | -e11 | -e10 | -e13 | -e12 | -e15 | -e14 | -e1 | -e0 | e3 | e2 | e5 | e4 | e7 | e6 |
e10 | e10 | -e11 | e8 | e9 | -e14 | -e15 | -e13 | -e12 | -e2 | -e3 | -e0 | e1 | e7 | e6 | e4 | e5 |
e11 | e11 | e10 | e9 | e8 | -e15 | -e14 | -e12 | -e13 | -e3 | -e2 | -e1 | -e0 | e6 | e7 | e5 | e4 |
e12 | e12 | -e13 | -e14 | -e15 | e8 | e9 | e11 | e10 | -e4 | -e5 | -e7 | -e6 | -e0 | e1 | e2 | e3 |
e13 | e13 | e12 | -e15 | -e14 | e9 | e8 | e10 | e11 | -e5 | -e4 | -e6 | -e7 | -e1 | -e0 | e3 | e2 |
e14 | e14 | -e15 | e12 | e13 | e10 | e11 | e8 | e9 | -e6 | -e7 | -e4 | -e5 | -e2 | -e3 | -e0 | e1 |
e15 | e15 | e14 | e13 | e12 | e11 | e10 | e9 | e8 | -e7 | -e6 | -e5 | -e4 | -e3 | -e2 | -e1 | -e0 |
Note that the 7 heptads have an element in common - in this case, e1. In general, any 2 heptads must intersect in a triad, any 3 heptads must intersect in at least a unit element and may intersect in a triad; if a system has 4 heptads whose common intersection is empty, then there must be 15 heptads in the system - i.e. in a system with exactly 7 heptads, those heptads must contain a common unit.
Suppose the contrary, that there are 4 heptads with h1∩h2∩h3∩h4=∅. Any triad intersects any heptad in at least one unit element, so a triad present in 3 out of the 4 heptads would intersect the 4th in an element, contradicting the assumed empty intersection. So for each of the 6 distinct triads tij=hi∩hj, there is another heptad not among the original 4 containing tij, for a total of at least 10 heptads, hence (see below) 15 distinct heptads.
There is an interesting representation of the Fano plane embedded in the heptads of these 7-heptad systems. The 7 triads containing the unit element common to all the heptads, and the heptads themselves, comprise the points and lines (or vice versa by duality). Any 2 triads of this triad set determine a unique heptad by multiplicative completion (and a unique 3rd triad of the set, contained in that unique heptad), and any 2 heptads determine a unique triad by intersection (and a unique third heptad containing that triad). Multiplication tables for triads and heptads are induced, which (sans signs) are the same as the octonionic triad groupings, as described further in the section below (Heptad Products).
Testing the 15! permutations of unit indices yields a group of 192 permutations leaving invariant the above triads and multiplication table.
Systems with 8-14 Heptads
For reasons analogous to those in Systems with 4, 5, or 6 Heptads, hypercomplex system S having 15 imaginary units and more than 7 heptads must have 15 heptads. There must exist among the >7 heptads some subset that completes to a "self-contained" set of 7 heptads as described above in Systems with 7 Heptads, with further heptad(s) not part of this subset. This further heptad intersects each of the 7 of the self-contained set in a triad having yet another multiplicatively closed heptad containing it. Again, there is a strong analogy to the extension via XOR or Cayley-Dickson construction of the octonions to the sedenions.
Systems with 15 Heptads
The sedenion families derived from the XOR and Cayley-Dickson constructions all possess 15 heptads, satisfying:
- any 2 intersecting triads generate a heptad
- each imaginary unit is in 7 heptads
- each triad is in 3 heptads
The proofs:
- each heptad contains 7 triads, hence 7×6/2 pairs of triads; since distinct heptads intersect in a single triad, any pair of triads is thus in only one heptad, and thus there are 15×(7×6/2)=315 distinct pairs in all the 15 heptads; any triad intersects 18 others, so there are a total of 35×18/2=315 pairs of intersecting triads; thus any pair of intersecting triads is contained in exactly one of the heptads
- each unit e is in 7 triads, so there are 7×6/2 pairs of triads intersecting in e; each such pair generates a heptad, but each such heptad is generated by 3 distinct such pairs, so there are (7×6/2)/3=7 distinct heptads containing e
- any triad intersects 18 others, and each such intersection generates a heptad; within such a heptad, the given triad intersects 6 others; hence there are 18/6=3 distinct heptads generated by the given triad and its set of intersecting triads, i.e. the given triad is in 3 distinct heptads
Testing the 15! permutations of the unit indices yields 20160 = 8!/2 permutations that leave the triad list and multiplication table invariant (up to signs). The multiplication table below will thus yield 2×15!/8! distinct variants (not counting further variants from sign assignment/triad handedness) from permuting the indices of the unit elements:
0: e1,e2,e3 | 1: e1,e4,e5 | 2: e1,e6,e7 | 3: e1,e8,e9 | 4: e1,e10,e11 |
5: e1,e12,e13 | 6: e1,e14,e15 | 7: e2,e4,e6 | 8: e2,e5,e7 | 9: e2,e8,e10 |
10: e2,e9,e11 | 11: e2,e12,e14 | 12: e2,e13,e15 | 13: e3,e4,e7 | 14: e3,e5,e6 |
15: e3,e8,e11 | 16: e3,e9,e10 | 17: e3,e12,e15 | 18: e3,e13,e14 | 19: e4,e8,e12 |
20: e4,e9,e13 | 21: e4,e10,e14 | 22: e4,e11,e15 | 23: e5,e8,e13 | 24: e5,e9,e12 |
25: e5,e10,e15 | 26: e5,e11,e14 | 27: e6,e8,e14 | 28: e6,e9,e15 | 29: e6,e10,e12 |
30: e6,e11,e13 | 31: e7,e8,e15 | 32: e7,e9,e14 | 33: e7,e10,e13 | 34: e7,e11,e12 |
| e0 | e1 | e2 | e3 | e4 | e5 | e6 | e7 | e8 | e9 | e10 | e11 | e12 | e13 | e14 | e15 |
e0 | e0 | e1 | e2 | e3 | e4 | e5 | e6 | e7 | e8 | e9 | e10 | e11 | e12 | e13 | e14 | e15 |
e1 | e1 | -e0 | e3 | -e2 | e5 | -e4 | e7 | -e6 | e9 | -e8 | e11 | -e10 | e13 | -e12 | e15 | -e14 |
e2 | e2 | -e3 | -e0 | e1 | e6 | e7 | -e4 | -e5 | e10 | e11 | -e8 | -e9 | e14 | e15 | -e12 | -e13 |
e3 | e3 | e2 | -e1 | -e0 | e7 | e6 | -e5 | -e4 | e11 | e10 | -e9 | -e8 | e15 | e14 | -e13 | -e12 |
e4 | e4 | -e5 | -e6 | -e7 | -e0 | e1 | e2 | e3 | e12 | e13 | e14 | e15 | -e8 | -e9 | -e10 | -e11 |
e5 | e5 | e4 | -e7 | -e6 | -e1 | -e0 | e3 | e2 | e13 | e12 | e15 | e14 | -e9 | -e8 | -e11 | -e10 |
e6 | e6 | -e7 | e4 | e5 | -e2 | -e3 | -e0 | e1 | e14 | e15 | e12 | e13 | -e10 | -e11 | -e8 | -e9 |
e7 | e7 | e6 | e5 | e4 | -e3 | -e2 | -e1 | -e0 | e15 | e14 | e13 | e12 | -e11 | -e10 | -e9 | -e8 |
e8 | e8 | -e9 | -e10 | -e11 | -e12 | -e13 | -e14 | -e15 | -e0 | e1 | e2 | e3 | e4 | e5 | e6 | e7 |
e9 | e9 | e8 | -e11 | -e10 | -e13 | -e12 | -e15 | -e14 | -e1 | -e0 | e3 | e2 | e5 | e4 | e7 | e6 |
e10 | e10 | -e11 | e8 | e9 | -e14 | -e15 | -e12 | -e13 | -e2 | -e3 | -e0 | e1 | e6 | e7 | e4 | e5 |
e11 | e11 | e10 | e9 | e8 | -e15 | -e14 | -e13 | -e12 | -e3 | -e2 | -e1 | -e0 | e7 | e6 | e5 | e4 |
e12 | e12 | -e13 | -e14 | -e15 | e8 | e9 | e10 | e11 | -e4 | -e5 | -e6 | -e7 | -e0 | e1 | e2 | e3 |
e13 | e13 | e12 | -e15 | -e14 | e9 | e8 | e11 | e10 | -e5 | -e4 | -e7 | -e6 | -e1 | -e0 | e3 | e2 |
e14 | e14 | -e15 | e12 | e13 | e10 | e11 | e8 | e9 | -e6 | -e7 | -e4 | -e5 | -e2 | -e3 | -e0 | e1 |
e15 | e15 | e14 | e13 | e12 | e11 | e10 | e9 | e8 | -e7 | -e6 | -e5 | -e4 | -e3 | -e2 | -e1 | -e0 |
Heptad Products