Open any gallery of seals a search for the word “sigil” hands back. The lines are confident, the captions are short, and not one caption says where the line came from. There is nothing in such a picture to argue with, because what you argue with is a construction, and a construction is what it never shows.
A built seal differs from a drawn one in a single property. That property is checkable: the built seal can be repeated. Given a grid of twenty-two points and a spelling, the line comes out by itself, the same for everyone who works from that spelling. The converse is worth more. A line that misses these points was traced by nobody; it was drawn by hand to taste. Seeing that takes a minute and no Hebrew beyond telling twenty-two letters apart.
Twenty-two points set down before us
The grid was not invented here, and it was not invented for sigils. It is the Rose Cross, the lamen of the Golden Dawn, which lies in the archive as a page of its own. That page holds what is known of its origin, along with the reservations about what could not be established. A drawing needs one thing from it. The twenty-two letters lie in three rings, and the place of each is fixed by a number, not by the hand that holds the pencil.
Three mothers inside, seven doubles in the middle ring, twelve simples outside. The division came out of Sefer Yetzirah and leaves no remainder, since three plus seven plus twelve makes twenty-two.
| Ring | Letters | From the centre | Step round the circle | Point of the letter |
|---|---|---|---|---|
| inner, three mothers | 3 | 48.6 to 84.24 | 120° | 66.42 |
| middle, seven doubles | 7 | 84.24 to 123.12 | 51.43° | 103.68 |
| outer, twelve simples | 12 | 123.12 to 162 | 30° | 142.56 |
The numbers are units of the drawing in which the Rose stands on its page. In centimetres they mean nothing. What means something is the ratios between them, which hold at any size the seal comes out. The point of a letter is taken halfway along the petal, midway between the inner and the outer edge of its ring. That is the only convention which had to be added to the Rose, because the document the figure is drawn from gives no sizes.
How the line is drawn
There is one rule, which unfolds into six steps, each of them done with hands and a pencil.
- The spelling is broken down into consonants. Vocalisation marks come off, because a vowel point has no petal.
- Final forms are reduced to the basic ones. There are twenty-two petals, and the Rose keeps no separate place for a final form.
- Each consonant is replaced by the point of its petal, and the order of points repeats the order of letters. No other rule applies, so no room is left for taste.
- The points are joined in order by one broken line, without lifting the hand.
- A circle goes on the first point, and on the last a crossbar at right angles to the final segment.
- If the line comes back to a letter already used, the return is marked with a loop. Two identical letters side by side get no loop, since the segment between them has zero length.
Here is the first shell of the archive, Thaumiel, in the vocalisation the canon holds. Seven letters give seven points and six segments, and the line touches six different petals. A loop stands on Aleph, because Aleph occurs twice in the spelling and the line comes back to it.
One shell, two spellings, two figures
The seventh shell is called Harab Serapel in the archive, while the Hebrew beside it comes from the system of 777: עֹרֵב זָרָק, Ravens of Dispersion. The choice is recorded as decision D2, which does not hide the losing side, because that side had time to spread across the live site before anyone took it down.
The vocalisation חֶרֶב שְׂרָפֵאל is not entered into the canon. The archive's canon, decision D2
On paper that editorial line stops being editorial. The argument begins at the first letter. The canonical form has ע, whose gematria is seventy, while the rejected one has ח, whose gematria is eight. Both stand in the outer ring of twelve, exactly opposite each other, half a circle apart, so the two lines set out from opposite edges of the Rose.
Both lines are built by the one rule, and both are continuous. The canonical spelling gives six letters, five distinct petals and five segments, while the rejected one gives eight letters, seven petals and seven segments. The figures share exactly one segment. At the second and third letter both forms give Resh and Bet, so the stretch between those petals coincides literally, and the rest diverges.
The length diverged as well. The canonical line fits into four and a half radii of the Rose and the rejected one into four and nine tenths, which is a difference the eye does not read. What the eye does read is something else. One figure has a knot on the left and a short stroke up to the right. The other drops a long way down to Lamed at the very bottom of the circle, and the first has no such descent.
A seal is taken off a spelling, not off a name
Out of the two drawings comes a statement worth keeping in hand while you look at somebody else's picture. The Rose does not know how a name sounds. It knows letters only, so two records that sound like one name to the ear give two figures, and both are lawful while each is built from its own record.
The caption “a seal of Harab Serapel” therefore does not fix the figure. What fixes it is a pair: the name, and the spelling the line was taken off. Until the spelling is named, any number of pictures can stand under such a caption, and nothing is left to choose between them. That is what people bring to pictures like these: is this one right or not. The question is put correctly, though one picture is not enough to answer it.
A built seal can be repeated.
A drawn one can only be traced over.
How to check somebody else's seal
The check asks for no faith in the archive and no knowledge of the language. What it needs is the Rose at hand and the spelling the picture is attached to.
- Count the vertices, both ends included. There must be as many as there are consonants in the spelling, and when two fall together, the shared vertex is obliged to carry a loop, or the count will not close.
- Lay the picture on the Rose so the rings coincide, then walk the line from the circle. Every vertex is obliged to land on a letter's point instead of resting beside it.
- Read the letters in the direction of travel. What comes out must be the spelling captioned under the picture, and where no spelling is captioned, read what did come out and judge by that.
- Check the ends. One circle, at the beginning; one crossbar, at the end, lying across the final segment and not along it.
- A refusal is a result too. Where the picture sits on no grid and no spelling is named under it, there is nothing to check, and the honest conclusion here is about trust in the page, not about the seal.
One vertex is spoiled here, the second in order. It left the middle ring for the outer one and came to rest between Zayin and Het, nearer to Het, though not on its point. Everything else stayed as it was, the figure still looks like a seal, the line is continuous, and the circle and crossbar are in place. The very first step of the check fells it, and fells it in silence, without any argument about meaning.
What the check does not do is worth stating on its own. It does not decide which of the two spellings is the better one, since that is an argument about sources and gets settled in the canon and not on paper. It says nothing about whether a seal works, because geometry makes no statement on the matter. It answers one question: was this line traced, or was it drawn.
What was chosen here, and what was found
The arrangement of the twenty-two letters across the rings is taken from the page of the Rose Cross. That page writes out what in this figure comes from the order's document, what the archive built up for itself, and where the sources disagree so far that the choice falls to us.
The point halfway along the petal is a convention of the archive. The circle at the start, the crossbar at the end and the loop of return are the archive's convention too, and a plain reservation belongs here. The archive has opened no primary source with an exact reference that describes those three marks, so no name will be attached to them on this page. The check loses nothing by that, since it rests on the vertex having to lie inside its own petal, while the marks only make the line readable.
Both spellings come from the archive's canon together with decision D2, and no Hebrew letter here is written from memory. A program assembles the drawings. It parses the petal geometry straight off the page of the Rose and takes the spellings from the canon. It then verifies that every vertex landed inside its own petal, that the circle stands at the start, and that the crossbar closes the line at a right angle. So that the checking cannot go blind, seven deliberately spoiled drawings are fed to it, and a run counts only when all seven are caught.
What comes next
The most useful thing to do with what you have read is to open the Rose Cross and walk a name you know across it. The petals answer to hover, each letter carries its tunnel with both of its ends, and the ring order is the one that the drawings above are built on.
The next issue stays with this subject and changes material: the shells themselves carry names that differ from author to author, and the tables argue with one another while rarely naming whom.