Codex Seraphinianus: Some Observations

They say that the text of the Codex Seraphinianus was never meant to mean anything; all the same, I mean to treat it here as if it was. Sounds crazy? I tried being sane once, and it nearly drove me mad. Anyway, don't blame me; it was Luigi Serafini who started it.

I don't own a copy of the Codex; I'm working from some notes I took in Michael Everson's library.

The Structure of the Text

The Codex consists of eleven chapters, each of which contains the following pages:

The whole book ends with the word A11P0C1a0F3F3 (presumably `The End').

Page Numbers

Page numbering is reset to 1 at the beginning of the sixth chapter; the codex is thus divided into two volumes of approximately equal size, without distinctive titles, but with lines to the effect of `End of A10 volume' and `End of B1 volume' at the end of each. There is some sort of index (under the header L0L0C4a0F1L0) to the second volume only.

The number system in which page numbers are written is, on the whole, base-21. This implies that a number is expressed as a sum of what I'll call scorones (i.e., twenty-ones) and ones. (You may think of scorone as an Italian-style augmentative of score or a contraction of score+one.) Suitable for counting on one's fingers, toes and nose or something. Telefol uses a base-27 system, and its speakers have the same anatomy as the rest of us, which is more than can be said of many kindreds in Serafini's world.

Now there are 21 digits (from 0 to 20), by means of which the quantity of ones is expressed. Not much is remarkable about them, except that 7 consists of 5 followed by a dot on the line; 5 itself looks like a wedge (on the relevance of which anon); the digit 16 is something (but not another digit) followed by a dot on the line and the digit 17 is something else followed by a raised dot.

The quantity of scorones is expressed in an opaque way (by which I mean that the number of pages in each of the two volumes runs to less than 9*21, and it does not appear possible to guess what happens afterwards). This involves the use of a vertical bar, a vee (that is, an upside-down 5), a dot on the line, a raised dot (here written as i, v, . and ', respectively, except when the dots form part of a digit), the digit 5 (the wedge) itself and, mirabile dictu, iteration of the digit that expresses the quantity of ones. To make it even more complicated, a quantity of scorones between 3 and 6 is expressed in one way with a quantity of ones up to 14 and in another with ones from 15 onwards. Further deviations occur if the quantity of ones is 5 or 7. (Some of those -- but not all -- can be attributed to the apparently regular substitution, probably for æsthetic reasons, of 5v for 55 as well as vv.)

In the table below a tilde indicates the digit corresponding to the quantity of ones.

Ones 0--4, 6, 8--14 5 7 15--20
0 ~ 5 7 ~
1 i~ i5 i7 i~
2 ii~ ii5 ii7 vi~
3 vii~ vii5 vii7 5vi~
4 5vii~ 5vii5 5vii7 i5vi~
5 5vii~~ 5vii5v 5vii7v i5vi~~
6 ~ii~~ .vii5v vii7v i~i~~
7 i~~~i i5v5v '5v5vi i~~~i
8 ii~~~i ii5v5v i'5v5vi ii~~~i

It gets totally out of hand in the ninth scorone:

There are larger numbers here and there in the book (each of the portraits in the chapter on history is accompanied by two numbers, which presumably stand for the years between which the person lived, ruled the country, or something), but those are written in a system of their own. And in many places in the text there are sequences of number symbols that don't form a legitimate number according to the rules stated above (a very common sequence is 22).

The use of iteration in the notation may be able to teach us a lesson. In the systems used for writing numbers in our world (Arabic, Roman etc.) the repetition of a symbol indicates that its numeric value somehow participates two or more times in the number. For example, in the number 33 the value of the symbol 3 appears twice (3*101+3*100=33); in the number XX the value of the symbol X (i.e., 10) also appears twice (10+10=20). But in a Serafinian number such as 5vii66 (5621=111) the repetition of 6 does not indicate two sixes; it merely signals that the six ones are added to five rather than four scorones (cf. 5vii6=4621=90). By analogy, one may suppose that the very frequent iteration of characters in titles and other words in majuscules also indicates something other than repetition of the corresponding phonetic value -- if indeed they have a phonetic value.

Back to Words

Which brings us back to the writing system. (I'm only discussing words written in majuscules here -- titles of chapters, sections, subsections and paragraphs, for the most part.) Several dozen different characters appear in them, far too many for the writing system to be an alphabet, and there are too many long words for it to be a syllabary. Some characters occur very many times, others only once or twice.

What is even more striking, however, is the tendency of the characters, even the less frequent ones, to reoccur within the same word or group of words (e.g., within the titles of the various subsections and paragraphs in a section). If a character occurs in a word at all, there's a good chance that it occurs there at least twice -- perhaps thrice in a row (which is next to unseen in any sort of phonetic writing system), up to six times altogether. It is as if the headers of most pages in an English book were such words as bookkeeper, googol, grammar, Ouagadougou and Wassamassaw.

On many occasions the titles of a section and its first subsection are cognate, though not derived from a common base in uniform ways. The titles of several other sections are derived from the sequences of the first characters of the titles of some or all of the corresponding subsections (usually, but not always, in the same order). (This is quite common in the first two chapters, but occurs only occasionally in the rest of the first volume, and not at all in the second.)

A meaning-oriented writing system? Or a philosophical language?

And to make it worse, otherwise distinct characters which share certain components or are mirror reflexions of one another seem to form similar patterns. But I'll say no more now.


Telefol counting starts with the fingers of the left hand (1 being the pinky), progresses from the thumb (5) to the wrist, lower arm, elbow, upper arm and shoulder (6--10), the side of the neck, ear and eye (11--13) and thence through the nose (14) and right eye (15) to the right pinky finger (27). The Telefol idea of a very large number is kakkat=14*27=378.

Created and maintained by Ivan A Derzhanski.

Last modified: 29 September 2004.