I've recently come across and started getting an interest into the Elian script, an alternative to the standard Latin script alphabets. Although writing in Elian script has shown to not improve my horrible handwriting1, the script itself and its composition rules allow for some very interesting graphical results which may even compare to the wonderful effect that can be achieved, for example, with the Arabic abjad.

The basics

I'm not to go into details on how to write Elian, because the web is full of better resources than I would ever write, but I will recall some of the essential details, which are useful to what I intend on writing next.

The principle behind the Elian script is very simple: take the standard hash/number sign or Tic-tac-toe grid, and assign a letter to each cell, in columns, starting from the bottom left. Since the grid has 9 places, this gives you the symbols for A to I. The next nine letters (J to R) are the same, but with one (any, but one) of the stems made longer, and the remaining eight (S to Z) are like the second-tier letters, with a dot or other similar marker added.

An example exposition of the Elian alphabet, with sharp corners and very regular traits, follows (read left to right, top to bottom as you would any other)2.

Elian alphabet

Of course, there is no need for the glyphs to be so regular in shape and size, nor for the corners to be so sharp. This is just practical when illustrating the alphabet and drawing the shapes by hand-coding an SVG.

The power of the Elian script comes largely from its weak composition rules, the lack of size constraints and from the shape flexibility of the tier II and tier III letters.

Composition and shaping

The composition rules are simply a local “left to right, top to bottom”, allowing to put characters on top of each other, or next to each other, or a combination of the two, as deemed aesthetically more pleasing. One of the main consequences of this is the lack of the concept of a baseline like in most other scripts.

Since there are no size constraints, ‘wrapping’ characters in other characters also comes quite natural, and the shape flexibility of the letters from J to Z can give spectacular effects.

To see this composition in action, let's look at my nickname, Oblomov. If I just take the SVG alphabet illustrated before and juxtapose the letters keeping their top alignment (that comes from the SVG coordinate system) I obtain:

Oblomov in Elian, try #1

But I can also imagine a bottom baseline, with ascendants and descendants, obtaining:

Oblomov in Elian, try #2

Much better, or at least more familiar to us, given how we're used to seeing text.

But we aren't even starting to scratch the surface of what can be done in Elian. The next step is to alter the shape of some of the tier II glyphs so that we can do some kerning, Elian-style:

Oblomov in Elian, try #3: kerning #1

By switching the shape of the letter ‘o’ around, we can interlace it with some of the following letters, and we can also nest the ‘b’ in the ‘l’.

But there's actually no reason why the ‘o’ should be shaped the same way throughout the text, or the stem lengths be consistent through different letters:

Oblomov in Elian, try #4: kerning #2

In this attempt, all three possible ‘o’ glyphs have been used, and the short/long stems are glyph-specific: the long stem of the ‘m’ is as long as the ‘a’ stems, and shorter than the short stem of the ‘l’. It's important to notice that the short stems of the ‘m’ could have been made longer, but this would have made it harder to tell the letter apart from a ‘d’, violating one of the rules about text legibility.

{ }


To a newcomer, the Elian script can be quite confusing, due to the high similarity of the glyphs. However, when thinking about it, this is really not different from the high symmetry found in our standard Latin script, with letter such as ‘d’, ‘b’, ‘p’ and ‘q’ obtained by mirroring and rotations, or the hand-written l resembling a taller e, and so on. { Ambigrams. }


The Elian script was mostly developed for English. Its simplicity in this regard, that greatly helps in the fancy composition that makes Elian beautiful, become limitations when looking at possible applications of the script to other languages.


First of all, Elian lacks case: there is no distinction between uppercase and lowercase letters; although this is a pretty common occurrence in Eastern scripts, most Western ones have developed mixed cases in the centuries, and have since codified rules for the use of the different cases. In German, for example, nouns are marked apart from other parts of speech by their uppercase initial.

Although case is not that important in many cases, there are even in English, situations where the initial uppercase can change the meaning of a word. It would be useful if Elian could develop a bicameral system like the other Western scripts.

It's interesting to note that casing cannot be reduced to a simple matter of size: as we've seen, letters can have very different sizes within the same word, and this is controlled more by overall aesthetics than other rules.

Proposal: outlining

First of all, it is important to remark that a bicameral system is not strictly necessary: Western scripts have done without until the Middle Ages or so. An important feature of Elian would be that the script remain mostly in a single case, with ‘uppercase’ (discussed momentarily) only used when deemed necessary. So we can forget about the abuse in capitalization for languages that don't actually need it (did I mention for example how I cringe at the so-called Title Case? All lowercase is fine.)

This being said, when uppercase letters want to be used, they can follow the same composition rules as standard Elian, with the exception that they are highlighted by outlining: the main shape will not be the stroke of the glyph, but its outline.

Since no Elian glyph intersects, outlining a glyph is trivial. Of course, this should be done by keeping the stroke width of the outline on par with the stroke width of the standard characters. Back to Oblomov, this time with an initial capital, the result would be something like the following (back to a more linear structure):

Oblomov in Elian, uppercase initial

Accents and diacritics

A second aspect showing how Elian was essentially developed for the English language is its lack of friendliness towards the huge variety of accents and diacritics that are present in most other languages based on the Latin script.

For accents, the author does describe how to use them, but it's undeniably true that their presence tends to ‘break’ the rhythm and flow of the script. For diacritics, the problem grows even worse: while they are seen as modified Latin letters by foreigners to the languages that uses them, their are usually considered independent ones by the native speakers of those languages. How should they be treated in Elian?

I'll go into more details about this issue when discussing the applicability of Elian to other alphabets, and on the topic of ligatures.


Finally, the Elian script by itself does not provide for a numeric system. The simplest approach in this case is to just use the so-called (Western) Arabic numerals we are used to, but it's obvious that their shape does not fit particularly well with the dynamic nature of the Elian script, and that a writing system for numerals more in line with it would be preferable.

There have been a number of proposals (see for example a relevant thread in the Elian script subreddit), some of which are interesting (such as the use of Roman numerals, but with the letters in Elian scripts) but not the most practical.

One of the proposals that caught my attention is to write numerals as letters with an additional mark. This is similar to the way numerals were written in ancient Greek, and also in the Arabic abjad before the introduction of what has become the more widespread Arabic numeral system.

The biggest issue I have with this approach is that the digit zero looks a little bit alien, and there is no clear connection between the digit value and its shape (which is there when you look at stylized forms of most Arabic numerals).

Proposals: tagging the circle

The solution I've been pondering is based on ‘tagging the circle’, where the circle refers to the Elian glyph for the letter ‘e’, which normally turns into a circle in handwritten Elian. Since the ‘circle’ has four sides (eh), it's possible for numerals to be identified by the number of tagged sides, or the tag position, or a number of variations thereof.

An interesting potential offered by these approaches is that they can be easily extended to bases other than 10, and particularly higher than 10, without using letters instead of digits (as it's normally done e.g. with the common base-16 using letters A to F as the digits after 9).

Elian numerals proposal #1: digits 0 to 9

In the first proposal tags cross the sides of the square, in clockwise order starting from the top, with each subsequent number adding a tag. It's therefore easy to read the number by counting the number of tags, but the writing starts to get expensive in terms of the number of traits.

There are two ways to simplify this: by moving the tag rather than adding it (so that, for example, number 2 still has a single tag, but on the second side), and by coalescing multiple tags on the same side into some form of squiggle. Both these approach are also more friendly to handwritten Elian, since they allow more continuous drawing.

Elian numerals proposal #2: digits 0 to f

The resulting numeral system (shown with rounded corners and caps) is much less baroque than the first one, making it easier to cover all of the 16 hexadecimal digits, for example. Of course, in the standard base-10 numbering system, one would only use the first 10 digits (0 to 9).

Identifying the digits is still very easy, since they go in groups of four. As a bonus, there's also a resemblance between the tag on the digit for 5 and the Roman numeral V.

Other solutions are also possible, of course. For example, tagging the corners instead of the sides, a single corner from 1 to 4, two corners for 5 to 8, and so on. But I'm not going to develop other alternatives for the moment, since I consider this last one quite satisfactory.

For multi-digit numbers, we are going to follow the usual (Western) convention of writing the most significant digits first; the mixed direction (left-to-right, top-to-bottom) of Elian script will also be allowed in multi-digit numbers. Consecutive digits will be allowed to share one side, since the reading order is sufficient to resolve ambiguity. When a common side would be tagged by both numerals, the tags are placed in clockwise order from the most to the least significant.

As an example, we will write the (base-10) number 19670224 (which if you want you can read as the date February 24, 1967).

Number 19670224 in Elian

Among the higlights, we remark the difference between the single double tag on the right side of the digit 6 and the double single tag in the common side of the final pair 24.

When handwriting numbers, the consecutive circles on which tags are written can be drawn with a single winding trait, sort of like a long (sideways or vertical) Arabic 8 with as many loops as there are digits.

Other alphabets

One of the upsides of the Elian script is that its simple geometrical generation rules can be applied to any alphabet that has up to 27 letters. It's funny that it falls short of one letter from the Arabic abjad (28 consonants), while it has one extra possible glyph over the standard Latin alphabet (26 letters).

For example, the principle behind Elian could be used for Greek, with e.g. the glyph used for the letter ‘c’ in the (standard) “Latin Elian” script being used for the letter ‘γ’ in the “Greek Elian” script.

The situation is a little more complex for alphabets based on the Latin script. These alphabets often have a few ‘missing’ letters from the complete 26-letter Latin script, and sometimes additional letters: while many of these can be composed as a standard Latin script glyph plus a diacritic, some others come from completely different scripts (e.g. the Thorn still in use in Icelandic).

In this case, it would be appropriate to preserve compatibility with the standard correspondence between Latin and Elian scripts, and devise appropriate solutions (such as the ligatures discussed momentarily) to replace the missing letter forms. In a few cases, exceptions can be made: for example, in Turkish, the difference between the letters ‘i’ and ‘ı’ can be rendered by using the ninth Elian glyph for ‘ı’ and adding a dot to achieve the letter ‘i’: there would be no confusion with the non-existent 27th glyph because of the matching length of the two stems (additionally, the dot could be added in a safe position near the corner to further prevent confusion).

Turkish dotless (right) and dotted (right) ı in Elian

A few examples of how to work diacritics into Elian are discussed in the next section.

Ligatures and other letter forms

For most purposes, ligatures are typographical solutions to the problem of glyph stems interfering with each other: hence the need for ligatures such as ff or fl or fi, as would happen for example with two consecutive letters ‘f’. In this sense, they are only geared towards improving the aesthetics of the line.

Other letter forms, such as æ or œ, were born as ligatures but have been standardized as actual letters in some alphabets to represent specific sounds or etymological spelling of particular words. In fact, the ‘w’ letter itself was born as a ligature (guess what for).

Finally, ligatures such as the ampersand (& from ‘et’) or the “at symbol” (@, from ‘at’), have entered common usage beyond their original role.

So how does Elian scripts deal with ligatures? I would disregard, at least for the time being, the idea of typographical ligatures in Elian: due to its sophisticated and unusual set of composition rules, the need for a ligature arises practically never, and other brilliant solutions are available to achieve aesthetically pleasing results.

A more interesting issue is that of ligatures as “native” letter forms in alphabets based off the Latin script, and that of other symbolic ligatures. Finally, we shall see how some ligatures can be used to replace letters from extended Latin alphabets (we will see the example for the thorn letter (þ), replaced by a ‘th’ ligature).

Possible candidates for an Elian ‘at symbol’ (@)

The at-symbol is the simplest ligature we can recreate in Elian. There are a number of possibilities, some of which are pictured. The dot position in the top rightmost solution is quite arbitrary, and could be moved to a number of places.

The key component to these ligatures in Elian is that each letter preserves its individuality, to prevent confusion when identifying the components. While with ‘at’ this is quite simple, achieving the same clarity with ‘et’ (&) is much more complicated, since almost any joining of the ‘e’ with a stem from the ‘t’ is likely to cause confusion, creating a glyph that could be read as ‘ns’ or ‘nu’ rather than ´et’.

Possible candidates for an Elian ampersand (&)

The number of solutions in this case is much more limited. The key point for a correct reading in this case is the stem length: to ensure that the last letter is a ‘t’, the long stem must be the one connected to the ‘e’. Observe that the solution still isn't perfect, though, since there is still no indication that the first ligature component is not an ‘n’.

I see no clear solution to this issue, except by abandoning the ligature approach and either using simple juxtaposition, or some other graphical trick such as glyph intersection. In fact, these other proposed approaches are the only way to clearly indicate the components when ligating the ‘e’ with a tier II letter, such as in the œ digraph; for tier I letters, the equal stem length composition rules can help in identifying the component, just like the dot in tier III.

Ligatures and pseudo-ligatures for æ (top), œ, & (bottom) in Elian

The weaker form of glyph intersection shown here is to join the letters in such a way that it's clear that a stem ends without merging with a stem of the next letter: this is easier when typesetting things e.g. the way I'm doing it in SVG, much harder to do in writing, as it effectively kills the possibility of writing these ligatures with a continuous trait. The result is also not particularly appealing.

To conclude this section, we give a short look at the last possible use of ligatures in Elian that we mentioned: replacing letters from extended Latin alphabets by ligatures of Elian glyphs representing a possible transcription of the matching sound. This is obviously not an universal solution, but it can be used for things such as ‘th’, ‘ch’, ‘sh’ etc. We will have a look at a way to compose ‘t’ and ‘h’ as a substitute for the letter thorn (þ).

Since the letters to be merged in this case are facing opposite directions, I believe that the cleanest solution to the ligature is ‘wall sharing’, a practice which the script authrix abandoned in the early stage of the development of Elian.

Elian letter thorn (þ)

In the proposed examples, the letters are clearly recognizable, and there is no room for ambiguity. As an additional bonus, the rightmost option even has a thorny look. It's easy to see how the same principle can be applied for the German ß, by considering it as an ‘ss’ or ‘sz’ ligature: in fact, this could solve the eternal dispute between the eszett and scharfes S by actually providing distinct glyphs for each.

Elian letter eszett or scharfes S (ß), as a ligature of sz (top five) or ss (bottom two)

As an ‘sz’ ligature, ß offers a number of interesting opportunities (only 5 possible variations are show in the previous picture). The number of ‘ss’ combinations is somewhat reduced, but a few interesting effects can still be achieved. In fact, in the ‘ss’ case, we could even propose a simple doubling of the dot (no extra lines), although that would be quite a stretch on legibility, unless the convention is universally adopted3.

It's probably impossible to dictate general rules for these ligatures, although it's easy to see how a similar principle to the one used for the þ or ß could be applied to e.g. a ch or sh ligature. Moreover, the choice of these ligatures is obviously language dependent.

For example, a French speaker that would like to write a ç without adding a cédille to the Elian ‘c’ glyph might want to look into a ‘cz’ ligature (referring to the origin of the symbol) or a ‘ce’ ligature (referring to the pronunciation, and mirroring how the sweet ‘g’ sound is enforced by adding an ‘e’). There's also a matter of practicality: a ‘cz’ ligature can be easily achieved with wall sharing, whereas a ‘ce’ ligature is extremely hard to achieve unambiguously.

Proposed solutions for an Elian ç: cz and ce ligatures

In the ‘ce’ case, the left-to-right, top-to-bottom reading order of Elian prevents mid-stem adjoining of the ‘c’ to the ‘e’, but it also helps reducing the ambiguity of the ligature. The solution remains sub-optimal.

If we are willing to live with the consequent ambiguity, the small Elian ‘e’ as a loop at the corner of a letter may become the standard for altering any letter of the alphabet, thereby producing, for example, the ä (as ‘ae’), ö (‘oe’) and ü (‘ue’) common in German, or the Norwegian ø (again, ‘oe’). A less ambiguous solution would be to not insist on ligatures, and turn the (typographicaly) smaller ‘e’ into some sort of diacritic, in a way that might resemble the use of the iota subscript in Greek.

Proposed solutions for Elian ä, ö, ü as ligatures for ae, oe, ue (leftmost), or using a small ‘e’ as diacritic (other columns).

  1. I've always tried to look for some scripts that were more ‘compatible’ with my horrible handwriting, and on which I could try to develop an actual calligraphy rather than my usual unintelligible cacography, but there goes. Not even Arabic could do, despite its right-to-left writing direction. ↩

  2. Implementation note: older versions of Gecko and WebKit have a bug that prevent external CSS stylesheets from being loaded by SVG images; if the rendering seems off, consider upgrading your browser; Also, due to a bug in Opera/Presto's SVG rendering, the dots in the SVGs had to be implemented not as zero-length paths but as paths with a very small, but non-zero, length. ↩

  3. At a certain point during the history of the script, Elian herself used to write doubled letters once and mark them with double dots. She then decided to discard this option. ↩