The idea of a phonemic writing system has a ton of logic to it. See the preface to the Rune School lessons for more on that.

But there are two major issues that phonemic systems run into.

1. How do you accomodate all dialects? Yes, a phonemic system allows for variations of sounds, but there are many dialects where differences stray further.

2. Users need to understand phonemes in order to use it. In order to understand phonemes, users need to gain some skill in listening and classifying sounds.

Issue 2 is very much related to issue 1. When someone is asking how to write a word in a phonemic system, you must ask them questions about how they say it. They often think they are saying it one way, but they are saying it another. You often advise them to write it how they “think” they’re saying it regardless of how they “really” say it.

I ran into these issues again and again with Rune School.

A solution appears

One day, someone in the Rune School discord server explained that an etymological system could be better for these issues.

I thought it was interesting, but I ultimately wrote it off. I went back to tinkering and tinkering and tinkering on the current Rune School system. At the time, I was researching new ways to make the system more intuitive and English-native. I was doing things like mapping out regular sound changes and assigning runes to connected sounds.

The more I worked on that, the more I realized that I was essentially re-inventing an etymological system from first principles. So I decided to revisit this person’s idea. They had created a little chart that mapped all vowel phonemes that existed in Old English to Latin letter spellings. “Hey wait, didn’t you say this wasn’t phonemic?” Yes, I did.

Really, all an etymological system does is rewind the clock as far back as we can reasonably go to create our phonemic foundation. Then we take a snapshot of that and keep it the same forever. “Why?” Two reasons.

1. Believe it or not, there are many English dialects still alive today that are offshoots of much older versions of English than General American or Received Pronounciation. In order to have one global writing standard for English, we need to go further back in time than we realize.

2. The period of time known as “Middle English” was wild, orthographically speaking. There were many inconsistencies or plain errors that caught on in this time that we can easily fix if we just go back a bit further.

Let’s look at an example. Take the words Sew, So, Sow.

All of these words in the Rune School system would be spelled with ᚩᚹ or ᚩ‍ᚹ. But not all dialects or accents today consider these words to be the same phoneme or lexical set. Mainstream accents like General American simply have the Tow-Toe Merger.

Some accents would put Sew and Sow in one category and So in another. Notice that these aren’t just different realizations of the same phoneme. They are two different categories entirely. Sure, Rune School could support spellings like ᚩᚩ for “Toe” words and ᚩᚹ for “Tow” words, but what about accents that don’t consider Sew and Sow to be in the same lexical set? Believe it or not, there are dialects used today that consider these two to be in different categories as well!

The easiest way to include all dialects is to simply go further back in time. And when we do this, we arrive at, essentially, an etymological spelling system.

• Sew was originally Siw    
• Sow was originally Soaw
• So was originally Swoa

English speakers are already accustomed to making accomodations in this way. We all use “what” even though neither General American nor Received Pronunciation pronounce that word with an “a” sound.

Arguing against myself

Isn’t “what” part of the issue I explained in the preface of Rune School?

Now, the original idea of an alphabet is simple; like a hammer. But today with English, the way that we use the alphabet (how we spell) has become something complicated; like a large factory machine for which we have lost the instructions.

80% of the words in this etymological system are not written super differently from our current “orthodox” spelling system. So what is required is just a re-framing of the value of English writing. If language is a technology to connect people together, then this system would be the most effective at doing that.

Is it more complicated? Perhaps slightly. But the benefit of potentially unifying all dialects together and increasing the connective power is worth it.

What about this next part in the preface?

Fix our current Latin alphabet spelling? The amount of proposals to do so are countless. Ultimately, it would take a united effort of all English-speaking nations to enforce such a proposal. Even if that happened, would the people accept it? When you spell words a certain way your whole life, the words are like childhood friends. A government requiring you to change your spelling later in life is like killing your old friends!

You could start using your preferred spelling change today, but it would be percieved by others as spelling mistakes. In order to be able to adopt a new spelling system and avoid the confusion with the current spelling system, a new alphabet (script) is needed.

I’m still correct about that. The difference is that because we are going back in time and correcting the errors that accrued in the past, the spellings in this system would be technically speaking more correct than the orthography that we have today.

The only problem is, if an alphabet can be created arbitrarily, it can be changed arbitrarily. It has no tie to history to make it feel concrete.

So it’s taking the historical rootedness argument from the preface to its fullest. And if language is meant to connect people together, going further into history will also help to recover our shared heritage as English speakers.

English is a blend of Germanic languages. Such a historic writing system would also serve to connect English with its own cousins, sisters, brothers, and ultimately its mother (Proto-Indo-European).

What we did

I want to just do things, so I took the mapping of OE phonemes to latin letter spellings and I started making a dictionary file. Myself and the creator of this system started from the most common words, working our way down.

We got ~1,000 words in the dictionary. I then exposed this dictionary via a translation tool so that we could generate sample sentences and get a feel for the system.

I then made a Keyman keyboard for typing and text prediction to help with spelling.

I then made a simple runic mapping to the latin letter spellings of this system. The runic mapping contains the same logic as the latin letter version, except it does not distinguish vowel length, which is accurate to how runes were used in history.

What this means is that the runic version of the etymological system actually ends up being quite succint. It becomes like more of a shorthand.

Examples

Modern Orthodox:
I know that the twain will never meet, but I am looking forward to the day next summer.

Etym Latin:
Ih cnoaw thæt the twœeyen will n·ea·f’r meet, b’·out ih ạm looc·ing fore·wạrd too the dæy neạh·’st sųmor.

Etym Runic:
ᛁᚻ᛫ᚳᚾᚪᚹ᛫ᚦᚫᛏ᛫ᚦᛖ᛫ᛏᚹᛟᚷᛖᚾ᛫ᚹᛁᛚᛚ᛫ᚾᚫᚠᚱ᛫ᛗᛖᛏ᛬ᛒᚢᛏ᛫ᛁᚻ᛫ᛠᛗ᛫ᛚᚩᛣᛁᛝ᛫ᚠᚩᚱᛖᚹᛠᚱᛞ᛫ᛏᚩ᛫ᚦᛖ᛫ᛞᚫᚷ᛫ᚾᛠᚻᛋᛏ᛫ᛋᚢᛗᚩᚱ᛫

Status

This is still very much early days. I have intentionally not shared the details of how this works yet. If you are interested in contributing, we want more people to ask questions and give feedback. So please come by the discord. The link is in the About page of this website.

Random thoughts

• A cool added benefit of an etymological spelling system is that if “spelling pronunciation” continues as a trend, English speakers could very well recover old pronunciations that used to exist. Some isolated people could just read a lot in this orthography and eventually their pronunciations would get much closer to the original Old English phonemes! Or at least the original “lexical sets” for words could be recovered.

• The Latin version encodes more information than the Runic version. For that reason, the Latin forms are helpful for learning and finding connections across pronunciations. It’s also just closer to what we are used to today. But the Runic being more succinct, it boils words down to their essential forms. Not “ᛋᛚᛖᛖᛈ” but “ᛋᛚᛖᛈ”. Not “ᛋᚾᚩᚪᚹ” but “ᛋᚾᚪᚹ”. These Runic forms could be useful almost as symbols; recognizable at a glance by native speakers.

• Even though the runic etymological spellings are more simple (no vowel length), it surprisingly doesn’t increase the amount of words that are spelled the same. This is due to the lexical set organization of Old English and also the fact that there’s a lot more letters that are silent today that weren’t silent in the past.

There are many genius choices that Read made. We can see them explained in that thread. There are also baffling choices (𐑙𐑣, 𐑘𐑢, 𐑕𐑟, 𐑺𐑻) on which we can only speculate.

Nonetheless, that thread inspired me to revisit my own attempt at creating a better Shavian that I have not yet publicized.

The second row is me freehand writing with a mouse the new script:

In 2022, a gentleman created the Brevian (and subsequent Anglovian) scripts. There are some great ideas in there (particularly the vowels). I am being clear that I have taken some ideas from this and used it as a jumping off point.

Now here are my goals:

1. Consistency with design rules (such as voiced/unvoiced, etc) so that the design feels transparent to the user
2. Stronger logic of the vowels (taking influence from Geoff Lindsey and others)
3. More connection to the history and development of the English language itself (in letter design but also in phonological relations)
4. More flexibility (multiple meanings)

Ultimately, what I seek to do could be illustrated with the following example from a Git workflow.

Vowels

Here are the vowels:

I’ve kept these largely the same as Brevian.

Right away, you’ll see a layout that has columns and rows. Each row represents a distinct vowel category. Each column represents a consistent change applied to that category. This means that FACE and CHOICE have something in common as well as LOT and CHOICE having something in common. The relations are transparent in that way.

The 2nd column is stretched horizontally which can give a very aesthetically satisfying “push and pull” to a text which better matches the prosody of how it is pronounced out loud. It provides breathing room and some breaking up of large chunks of letters.

The 3rd column is the addition of the KIT vowel.

The 4th column is the addition of the FOOT vowel.

Consonants

Here are the consonants in the system:

I have kept these the same as Shavian and only changed what I deemed necessary to change.

In order, you can see the changes:

/s z θ ð f v n ŋ m r l h y w ʍ/

R and L

R and L are a sort of pair, just like in Shavian. The difference is that I made the shapes inspired by latin in order to avoid confusion with the new vowel shapes (LOT and DRESS).

The bonus for R is that it can still easily be attached on the ends of vowels just like in Shavian.

N and M

I knew right away that I wanted to connect /n/ with /ŋ/. This required a change in the design of N. I dug into the history of the latin letters and discovered the gorgeous N and M letters from Old Italic. A connection to history is always satisfying.

Y and W and friends

Connecting /j/ with the KIT vowel and /w/ with the FOOT vowel was another stroke of genius from Brevian.

Since I freed /h/ from being a pair with /ŋ/, I was able to simply adopt the lowercase latin h which fit quite well. It also created a beautiful combination letter for the famous /hw~ʍ/ sound in older English accents.

Historic Allophones

Historically, the following sounds are allophones in English: /sz θð fv/

You’ll notice that S and Z and the TH sounds look exactly the same except ascending and descending. Technically speaking, all of these consonant pairs are correctly rotated 180 degrees just like all the other non-allophonic pairs. So there is consistency in that regard. It just so happens that the design of the letters make SZ and the TH sounds look pretty much the same either way around!

This is intentional. S and Z still remain very close allophones even in the modern day (think of “dogs” vs “cats”, etc), as do θ and ð (think of “the” or “with”).

The letters for /f/ and /v/ are inspired by the Greek phi. Now you might say “But F and V were historic allophones just like those others! Why do these letters look more different when rotated?”

That is because F and V split to become distinct phonemes much earlier than the others.

As I understand it, the order of strongest allophone pair to weakest today is /θð/, then /sz/, then /fv/.

/n/ can also be an allophone pair with /ŋ/, which is why I wanted them to look more similar. But they’re not exactly a pair in the same way that the others are.

Future Potential

Here are some areas for consideration regarding future development. I am open to collaboration and feedback.

Descenders on certain vowels

I considered adding descenders on both the 3rd and 4th columns of the vowel chart. This would more explicitly show the addition / combination of /j/ and /w/ onto the initial vowel rather than KIT and FOOT. This could be a dialectal preference or useful in another way.

Dotted vowels

In addition to the the 2nd column in the vowel chart being stretched lengthwise, I also considered a layout that used a dotted motif instead.

This is pretty much up to your subjective taste. On one hand, I like the aesthetic of the variable width of vowels that more accurately match the (sometimes) variable length of vowels when spoken. On the other hand, the overdots could serve to bring the script more “up”, balancing out the plentiful descenders.

One option is to save either the stretched or the dotted column for an extra column that would only be needed in certain scenarios. For example, if a speaker distinguishes BAD from LAD, they could use this extra “column” as the motif for their vowel.

Etymological spelling?

This is a big, big question. Recently, I had been discussing with some colleagues the idea of an etymological spelling as opposed to a phonemic spelling. This would of course be against the explicit goals of George Bernard Shaw (Shavian), but the benefit is that it would absolutely ensure unity among all English dialects / accents in a way that a phonemic script could not.

This is where I realized that some of these extra ideas I had for vowels could come in handy. I will leave the below image without explanation as a sort of “if you know, you know” thing.

I’m not sure if this is worth pursuing or not, but I can sleep at night knowing that the system has the potential to at least make it possible.

Fin

Come stop by the Rune School discord if you would like to discuss this post.

1. Younger Futhorc
2. American Futharch
3. Futhorc.com
4. Rune Revival
5. Rune School (here)

I will include a first pass at a compromise as I outline each.

Consonants

Vowels

Now let’s look at how the vowel runes land on the IPA vowel chart for a stronger visual comparison.

Younger Futhorc:

American Futharch:

Futhorc.com:

Rune Revival:

Rune School:

Want to try your hand at a layout? Give it a shot!

Analysis

Now just comparing sound assignments doesn’t fully explain the reasons why the creators of these systems made the choices they did. Each sound assignment has ripple effects that will change the sound assignments for other runes.

In order to really compare these choices, we need to talk about the underlying values and philosophy. For example, are we going for authenticity? Are we going for clarity of design?

I would say that American Futharch aims to be more opinionated, as a work of art. Younger Futhorc and Rune Revival aim to be more authentic. Futhorc.com and Rune School seem to aim for authenticity while also wanting something more systematic. For example, Futhorc.com and Rune School both shared the vowel+semivowel digraph for the diphthongs. And Futhorc.com even goes further and uses t+sh for the “ch” sound (very analytical).

One of the biggest decisions is whether or not to represent the commA sound (schwa). Younger Futhorc opts to not represent it. The tradeoff then is you need to be comfortable with 5 ways of writing a schwa and 3 ways of writing the NURSE vowel. It’s not terrible, but you still completely lose the sound-spelling connection.

Shared values

My primary goals for a good modern rune system would be:

1. Authenticity to the original Futhorc

I like one principle that Rune Revival adheres to. They are willing to draw from the Elder Futhark, but they won’t draw from any traditions later than the Anglo Futhorc. I find that to be an authentic compromise when thinking outside the box is needed. Another aspect of authenticity is the most common runes still being the most common in a modern system (if possible).

1. Usable for all English accents

A couple of these systems are fairly obviously designed for American English. I think a good system would work for the entire English speaking world. Because the Futhorc is the “birthright” of English everywhere.

1. Systematically satisfying

Regarding Systematics, I am speaking mostly about phonemic mergers and splits as well as historical development of sounds. Things like the FOOT-STRUT split, the COT-CAUGHT merger, etc. A good system should take these into account.

Compromise

Now what would a compromise between all of these systems look like? The authenticity of the Younger Futhorc but with an added element of a more robust… system?

After using and teaching the Rune School system, I can say that it really does feel intuitive, simple, and functional. This is because time was spent thinking about how to phonemically analyze English as a complete system. There are 7 base vowels, from which all other vowel sounds come.

What I am looking for in improvements to the Rune School system:

1. Being more functional for people with different accents

2. More authentic to the original Futhorc

Let’s try something as a way to start the conversation:

This draws on the Elder Futhark for the assignment of ᛇ. This gives us an immense benefit to representing “hell’s corner” with all similar looking runes. It also lets us distinguish TOE from TOW, PAIN from PANE, and potentially rosEs and happY. The only thing that is lost is potentially a cleanness of the system as a whole. For example, lettER is no longer grouped with commA and contrasted with NURSE-STRUT. The 7 base vowels of Rune School would be upped to 9 base vowels. And such a system has not taken into account the frequency of phonemes, either. In the original Futhorc corpus, ᚢ was extremely common, which may lead one to use ᚢ as a commA (schwa) phoneme instead. Or even merge FOOT with commA (there are no minimal pairs between these two, so it’s possible, technically).

If you designed one of these systems, or if you have your own that I have not included, please join the Rune School Discord server and we can discuss in a sort of “peer review” working group. I’ll see you there!

• ᚫ turned into ᚫ ᚪ ᚩ
• ᚷ turned into ᚷ ᚸ
• ᚳ turned into ᚳ ᛣ

One rune will split into multiple different forms when that one rune begins to take on new sounds. When new distinctions within the sound become more apparent and meaningful, a new rune will be made. At such a point, you have two (or more) runes in the “set”.

The most interesting thing about this is that the oldest, original rune in the set doesn’t keep the oldest, original sound.

/g/ and /j/

An example of this is when the ᚷ rune used to primarily make a “g” sound like in the word “gift”. Secondarily, it made a “y” sound like in the word “yell” or “yule”.

So over time, we had a scenario like the above where one rune took on new sounds and the two sounds became more and more distinct. Some rune masters (whoever they were), then made a new ᚸ rune for one of these sounds.

Now any person would naturally assume that a new rune would get assigned the new sound (the “y” sound in this case). But no, the new rune got the older sound (the “g” sound).

The same thing happened to ᚳ and ᛣ. The primary sound of ᚳ was “c” as in “cat” and secondarily “ch” as in “chat”. The new rune ᛣ was made and it got the old sound of “cat”.

Laws of Analogy

Why is this the case? I was pointed to a man named Jerzy Kuryłowicz and his Laws of Analogy that he developed in regards to how analogical grammatical changes work in Indo-European languages.

The laws are as follows:

1. A bipartite marker tends to replace an isofunctional simple marker.
2. The directionality of analogy is from a “basic” form to a “subordinate” form with respect to their spheres of usage.
3. A structure consisting of a basic and a subordinate member serves as a foundation for a basic member which is isofunctional but isolated.
4. When the old (non-analogical) form and the new (analogical) form are both in use, the former remains in secondary function and the latter takes the basic function.
5. A more marginal distinction is eliminated for the benefit of a more significant distinction.
6. A base in analogy may belong to a prestige dialect affecting the form of a dialect imitating it.

In simpler English (thanks to ChatGPT v4):

1. A two-part grammatical marker (like “will have” for future perfect tense in English) often replaces a one-part marker that does the same job (if it existed).
2. Changes in grammar tend to spread from the more commonly used form to the less commonly used form. For example, a common word might influence the form of a less common word.
3. A grammatical structure that has a main part and a secondary part can lead to changes in another similar but standalone structure. For example, a verb form that usually comes with an auxiliary verb might influence a similar verb form that doesn’t use an auxiliary.
4. When both the old and the new versions of a word or grammatical form are used, the old one gets a less important role, and the new one becomes the main version. For example, “dove” as a past tense of “dive” became more common than the older “dived.”
5. Less important grammatical differences tend to disappear in favor of more important ones. For example, minor pronunciation differences in word endings might be dropped to keep more important grammatical distinctions clear.
6. The way people speak in a prestigious or standard dialect can influence how people speak in other dialects. For example, regional dialects might adopt grammatical forms from the standard language used in formal settings.

These laws are related to grammar, but if they successfully describe grammar, the chances are high that they can be pretty well abstracted to other things as well. (without getting too off-topic, I would say that life itself is grammar.)

For example, runes evolved exactly as law number 4 says they would. The older form (ᚷ and ᚳ) remained in secondary function and the newer form (ᚸ and ᛣ) took the basic function.

Application to Rune School

Now lets go through the laws and see how they might apply to our thinking about the Rune School system in other ways.

A two-part marker tends to replace a one-part simple marker.

In English with Latin letters, we have a number of ways to spell the GOAT vowel. In some words, they just use a single “o” letter for this sound. Eg. “no”, “so”, “go”

There are also words that use two letters like “ow” for this sound. Eg. “row”, “mow”

Rune School uses the two-part marker for this GOAT vowel ᚩᚹ which then replaces any one-part marker. Seemingly, humans just prefer composability and that is why things will trend towards being composable if at all possible.

The directionality of analogy is from a “basic” form to a “subordinate” form with respect to their spheres of usage.

As mentioned, we use the form ᚩᚹ for the GOAT vowel, which is basically “ow” in English with Latin letters. This is like the words “show”, “snow”, “bowl”, “low”, “grow”, “flow”.

We could then say that these very common “ow” words influence the less common words. So we use an “ow” form (ᚩᚹ) for all similar vowel sounds. The “ow” form has replaced the less common “-ough”, “-ew”, “-eau” words.

A structure consisting of a basic and a subordinate member serves as a foundation for a basic member which is isofunctional but isolated.

By making heavier use of two-part rune combinations, the Rune School system helps to reinforce the understanding of individual runes.

For example, by using ᚫᛡ for the sound in “price”, we are creating a more robust foundation for the individual ᚫ rune. If you pronounce “price” like [æi̯], [ai̯], or [ɑi̯], you then learn that ᚫ could sound like [æ], [a], or [ɑ]. This is important because in a phonemic system, individual letters need to feel a bit malleable in that way.

When the old (non-analogical) form and the new (analogical) form are both in use, the former remains in secondary function and the latter takes the basic function.

English used to have a main “u” sound which was like the word “goose”. Then this sound shortened to be like “foot”, at which point we had two sounds in the “u” category. Then the “foot” sound split into both “foot” and “strut”.

The idea is that the ᚢ rune used to have the “goose” and “foot” sounds and then when the sound split into “strut”, the newer ᚣ rune took on the older sound. So ᚢ is then reserved for the secondary function of “strut” (even if “strut” is quite a common sound in modern English).

Similarly, in Rune School we use ᚷ for the J sound in “joke”. It’s no secret that in English the “soft j” sound like in “yule” slowly became a “hard j” like in “John”. So as ᚷ changed (yet again), the newer rune ᛡ took on the older sound. So ᚷ is reserved for the secondary function of “Jello” while ᛡ has the primary function of “yellow”.

So ᚢ and ᚷ follow the same pattern of ᚫ and ᚳ.

A more marginal distinction is eliminated for the benefit of a more significant distinction.

In the standard Rune School system, we follow Shavian’s lead in eliminating as much as possible the distinctions in sound that just don’t matter all that much. The benefit of this is that comprehension becomes easier

For example, we only have 7 rows in our vowel chart.

We could have added another row specifically for ᚪ (/ɑ/-ish sounds), but by doing so, the distinctions between sounds become smaller. So instead, we make ᚪ a shortcut for writing ᚫᚫ.

We eliminate the more marginal distinction between ᚫ and ᚪ for the benefit of a more significant distinction between ᚫ and ᚩ.

The way people speak in a prestigious or standard dialect can influence how people speak in other dialects. For example, regional dialects might adopt grammatical forms from the standard language used in formal settings.

It’s no secret that Rune School uses a somewhat “fancy” pronunciation for a standard. It follows Shavian’s lead in this regard.

But according to this law, this is probably a good thing, seeing that the “prestige” dialect tends to influence the others.

What do you think? Let us know.

You may notice that the look of Rune School has changed. This is because we have swapped the primary font to Catrinity!

Catrinity recently updated to version 2.09 which adds a host of new bindrunes, among other things. Now instead of writing “Ohio” like ᚩᚹᚻᚫᛡᚩᚹ we can write it like ᚩ‍ᚹᚻᚫ‍ᛡᚩ‍ᚹ. And since we have a bind for ᛋᚳ, it will be more clear that we are writing “pasture” (ᛈᚫᛋᚳᛟᚱ) instead of something like “passion” (ᛈᚫᛋ‍ᚳᛟᚾ).

New Dictionary

Rune School offers a dictionary that should help new learners with the spelling standard. There is also a naive auto-translation tool at your disposal.

Word Game

If you’re looking for more immersion in runic spelling, see the new daily word game! All of the instructions are even in runes.

The word game uses the new dictionary, and thus also uses the spelling standard. But it does not use some shortcut runes in order to make the game slightly easier.

• ᚪ -> ᚫᚫ
• ᛇ -> ᛁᛡ
• ᛠ -> ᛁᛁ
• ᛥ -> ᛋᛏ
• ᛢ -> ᛣᚹ

There is a channel on the Rune School discord server where you can post your daily results and compare with others.

Now, which are the most common runes in our Rune School Spelling System? There’s two different ways we can determine this. The first is an analysis using the Shavian Readlex dictionary data. The second is by calculating frequency based on some transliterated sample text.

Dictionary method

Since the spelling is largely based on the Shavian ReadLex standard, we should be able to use that as a base of information to determine this.

@kj7qlv in the Rune School discord server wrote a script that analyzes the IPA spelling in the Shavian ReadLex dictionary directly. This way we can get the data on things like our “happY” rune ᛄ.

import pandas as pd
import re
import argparse

# Setup argument parser to accept flags
parser = argparse.ArgumentParser(description='Calculate the most common runes for the Rune School Spelling System.')
parser.add_argument('--weighted', action='store_true', help='Apply weighting based on frequency')
parser.add_argument('--shortcuts', action='store_true', help='Use shortcut runes ᛇᛠᚪ')
parser.add_argument('--v', action='store_true', help='Use the Double Feoh bindrune for /v/ phoneme')
parser.add_argument('--kw', action='store_true', help='Use Cweorth rune for /kw/ phoneme')
parser.add_argument('--st', action='store_true', help='(Naively) Use Stan rune for /st/ phoneme')

# Parse the command line arguments
args = parser.parse_args()

# Sample regular expressions and phonemes data
phonemes = {
    re.compile(r'(?<![aeɔ])ɪ'): {'ᛁ': 1},
    re.compile(r'i(?!ː)'): {'ᛄ': 1},
    re.compile(r'e(?!ə)'): {'ᛖ': 1},
    re.compile(r'æ'): {'ᚫ': 1},
    re.compile(r'ɒ(?!ː)'): {'ᚩ': 1},
    re.compile(r'ɔːR'): {'ᚩ': 1, 'ᚱ': 1},
    re.compile(r'ʊ'): {'ᚣ': 1},
    re.compile(r'ʊəR'): {'ᚣ': 1, 'ᚱ': 1},
    re.compile(r'ʌ'): {'ᚢ': 1},
    re.compile(r'ɜːR'): {'ᚢ': 1, 'ᚱ': 1},
    re.compile(r'ə(?!R)|I'): {'ᛟ': 1},
    re.compile(r'əR'): {'ᛟ': 1, 'ᚱ': 1},
    re.compile(r'iə(?!R)'): {'ᛠ': 1} if args.shortcuts else {'ᛁ': 2},
    re.compile(r'iəR'): {'ᛠ': 1, 'ᚱ': 1} if args.shortcuts else {'ᛁ': 2, 'ᚱ': 1},
    re.compile(r'eəR'): {'ᛖ': 1, 'ᚱ': 1},
    re.compile(r'(ɑː|Ɑ)(?!R)'): {'ᚪ': 1} if args.shortcuts else {'ᚫ': 2},
    re.compile(r'(ɑː|Ɑ)R'): {'ᚪ': 1, 'ᚱ': 1} if args.shortcuts else {'ᚫ': 2, 'ᚱ': 1},
    re.compile(r'ɔː'): {'ᚩ': 2},
    re.compile(r'iː'): {'ᛇ': 1} if args.shortcuts else {'ᛁ': 1, 'ᛡ': 1},
    re.compile(r'eɪ'): {'ᛖ': 1, 'ᛡ': 1},
    re.compile(r'aɪ'): {'ᚫ': 1, 'ᛡ': 1},
    re.compile(r'ɔɪ'): {'ᚩ': 1, 'ᛡ': 1},
    re.compile(r'aʊ'): {'ᚫ': 1, 'ᚹ': 1},
    re.compile(r'əʊ'): {'ᚩ': 1, 'ᚹ': 1},
    re.compile(r'uː'): {'ᚣ': 1, 'ᚹ': 1},
    re.compile(r'p'): {'ᛈ': 1},
    re.compile(r'b'): {'ᛒ': 1},
    re.compile(r'st'): {'ᛥ': 1} if args.st else {'ᛋ': 1, 'ᛏ': 1},
    re.compile(r't(?!ʃ)'): {'ᛏ': 1},
    re.compile(r'd'): {'ᛞ': 1},
    re.compile(r'kw'): {'ᛢ': 1} if args.kw else {'ᛣ': 1, 'ᚹ': 1},
    re.compile(r'k'): {'ᛣ': 1},
    re.compile(r'(ɡ|g)'): {'ᚸ': 1},
    re.compile(r'f'): {'ᚠ': 1},
    re.compile(r'v'): {'v': 1} if args.v else {'ᚠ': 1},
    re.compile(r'(θ|ð|Ð)'): {'ᚦ': 1},
    re.compile(r's'): {'ᛋ': 1},
    re.compile(r'z'): {'ᛉ': 1},
    re.compile(r'(?<!t)ʃ'): {'ᛋ': 1, 'ᚳ': 1},
    re.compile(r'(?<!d)ʒ'): {'ᛉ': 1, 'ᚳ': 1},
    re.compile(r'tʃ'): {'ᚳ': 1},
    re.compile(r'dʒ'): {'ᚷ': 1},
    re.compile(r'j'): {'ᛡ': 1},
    re.compile(r'w'): {'ᚹ': 1},
    re.compile(r'ŋ'): {'ᛝ': 1},
    re.compile(r'h'): {'ᚻ': 1},
    re.compile(r'l'): {'ᛚ': 1},
    re.compile(r'r'): {'ᚱ': 1},
    re.compile(r'm'): {'ᛗ': 1},
    re.compile(r'n'): {'ᚾ': 1}
}

# Initialize running_total dictionary
running_total = {}

# Function to update running_total based on phonemes match
def update_running_total(phoneme_dict, frequency, weighted: bool):
    for key, value in phoneme_dict.items():
        value = value * frequency if weighted else value
        if key in running_total:
            running_total[key] += value
        else:
            running_total[key] = value

# Function to initialize readlex DataFrame from TSV file
def initialize_readlex(file_path):
    try:
        readlex_df = pd.read_csv(file_path, sep='\t', header=None, usecols=[3, 4])
        readlex_df.columns = ['Pronunciation', 'Frequency']
        return readlex_df
    except Exception as e:
        print(f"Error occurred while initializing readlex: {e}")
        return None

# Sample data for testing
readlex_df = initialize_readlex('./kingsleyreadlexicon.tsv')
if readlex_df is not None:
    # Iterate through readlex
    for index, row in readlex_df.iterrows():
        pronunciation = row['Pronunciation']
        frequency = row['Frequency']
        
        # Check against all keys in phonemes
        for regex, phoneme_dict in phonemes.items():
            matches = regex.findall(pronunciation)
            for match in matches:
                update_running_total(phoneme_dict, frequency, weighted=args.weighted)

running_total_sum = sum(running_total.values())
for key, value in running_total.items():
    running_total[key] = value / running_total_sum
running_total_sorted = dict(sorted(running_total.items(), key=lambda item: item[1], reverse=True))
print(running_total_sorted)

If you put the kingsleyreadlexicon.tsv dictionary in the same directory as this script and do the command python script.py --help you will see the following:

usage: script.py [-h] [--weighted] [--v] [--shortcuts] [--kw] [--st]

Calculate the most common runes for the Rune School Spelling System.

options:
  -h, --help   show this help message and exit
  --weighted   Apply weighting based on frequency
  --shortcuts  Use shortcut runes ᛇᛠᚪ
  --v          Use the Double Feoh bindrune for /v/ phoneme
  --kw         Use Cweorth rune for /kw/ phoneme
  --st         (Naively) Use Stan rune for /st/ phoneme

Results

So here is the data with all of the runes except for ᛥ and ᛢ.

Unweighted means that the frequency of the word doesn’t matter.

python script.py --v --shortcuts

Weighted means that the frequency of the word is taken into account. This data will reflect how often you will see runes in a novel, for example.

python script.py --weighted --v --shortcuts

You can compare similar analysis done in Shavian.

Sample text method

Now lets look at the data from running some sample text. I have transliterated a decent amount of text from some Ernest Hemingway stories because he wrote in a very simple manner, which is good if we want to get a sense for rune frequency of everyday writing.

And here is an analysis without the shortcut runes ᚪᛇᛠᛥᛢ.

Findings

Now lets compare these two results with the original Futhorc corpus frequency:

Naively, we can just sum up the points of change as being 220.

How much of these changes are just due to the English language itself changing? It’s hard to say, but I suspect a lot.

Scrabble

If we wanted to get data for a game like Scrabble, we could get rid of the shortcut runes (ᛇᛠᚪ) and only consider the unweighted frequency. This would ensure a much more even frequency distribution, which I think is what we would want for this game.

python script.py --v

Now to view this data in a Scrabble Letter Distribution table, we can assign tile numbers and point values like so:

ᛒᚫ‍ᛡ᛫​ᚱᚩᛒᛟᚱᛏ᛫​ᚠᚱᚩᛥ

Stopping by Woods on a Snowy Evening

by Robert Frost

Let’s quantify this in terms of strokes per word.

There’s a few things to note in this simple test.

1. Shavian is much easier than runes
2. Runes are similar to writing all uppercase Latin letters
3. Chinese is much harder than runes

Let’s think about the place of handwriting in Chinese society. Other than their own names, most adults in China do not write Chinese characters on a day-to-day basis. It’s true that in school, the Chinese write characters by hand. But this is mostly for memorization purposes. Today, most people just type Chinese characters one way or another.

Now let’s think about how technology has changed our communication priorities over time.

From the 1900’s to the 2000’s, our communication technology and preferences have seemingly completely inverted.

As our technology progresses, we are returning more and more to a pre-modern mode of communication. The difference is that we are able to fully simulate a face-to-face communication across distance rather than requiring travel.

Needless to say, the place of handwriting has changed. Where it was once an essential tool for communication across distance, today it has been relegated to a kind of strange, forgotten middle ground. When you do write, you are typing it on a screen. When you write something by hand, it naturally takes on a special kind of significance because it is not the norm. And when something is special, it’s okay if it requires more effort.

But what if the power goes out? What if you want to unplug? If you wanted to reduce the amount of electronics in your life, Shavian would certainly assist you there. It is a tool for incredible efficiency of writing by hand.

Perhaps in the 2100’s, people will embrace actually writing by hand and seek a solution more like Shavian for that purpose. Or perhaps the advanced technology will continue to make handwriting mostly irrelevant. It’s easy for me to imagine a future in which technology enables us to mostly forget about the written word. As scary as it sounds, one can almost survive this way today.

• Books -> Podcasts and videos
• Letters -> Voice messages
• Searching for information -> Conversational AI assistant

The written word could essentially become a machine-only tool with an elite group of people that are literate just for the purposes of maintaining the machines.

As of today, handwriting remains of very little relevance to me. Runes may strike a nice middle ground between being just laborious enough to be special, and not quite laborious enough to be painful.

What do you think? Join the Rune School Discord server and let me know!

16EB	᛫	RUNIC SINGLE PUNCTUATION
16EC	᛬	RUNIC MULTIPLE PUNCTUATION
16ED	᛭	RUNIC CROSS PUNCTUATION

So working within these bounds, what can we do? Jump to the summary to see what I propose.

Less is more

Cormac McCarthy has 3 rules of punctuation which basically amounts to:

1. No quotation marks
2. No semicolons, and rarely colons.
3. I believe in periods, in capitals, in the occasional comma, and that’s it.

McCarthy also said:

James Joyce is a good model for punctuation. He keeps it to an absolute minimum. There’s no reason to blot the page up with weird little marks. I mean, if you write properly you shouldn’t have to punctuate.

To me, what McCarthy and Joyce are getting at is really the purpose of writing to begin with. We don’t hear punctuation most of the time. We certainly don’t hear quotation marks. At best we have short pauses and long pauses.

Full stop, comma, space

With this in mind, we can design some very basic principles with the material that we have.

Firstly, we have the space ᛫. That’s easy.

ᚦ᛫​ᛣᚫᛏ᛫​ᛁᚾ᛫​ᚦ᛫​ᚻᚫᛏ

The cat in the hat

A period or full stop is the longest pause that we have in speech. Visually, this should be reflected as the largest gap between runes. For this reason, let’s represent this as two of the space runes side by side ᛫᛫.

ᚦ᛫​ᛣᚫᛏ᛫​ᚹᛖᚾᛏ᛫​ᚻᚩᚹᛞ᛫​᛫​ᚫᛡ᛫​ᛏᚣᚹ᛫​᛫​

The cat went home. I went too.

Now McCarthy also uses commas. A comma is really just a smaller pause. So to represent this, we can turn the double dot full stop vertical ᛬. This way, it visually takes up a smaller gap between runes than the full stop, but it also shows something more significant than just a space.

ᚫᛡ᛫​ᛚᛖᚠᛏ᛬ᛒᚢᛏ᛫​ᚫᛡ᛫​ᛞᛁᛞᚾᛏ᛫​ᚹᚩᚾᛏ᛫​ᛏ᛫​᛫​

I left, but I didn’t want to.

Capitals

What stood out to me above however was McCarthy’s use of capitals. I wouldn’t normally connect the concepts of capitals and punctuation, so I asked Meta.ai to explain how he uses them

• No quotation marks: McCarthy often omits quotation marks to indicate dialogue, instead using a space or a line break to separate spoken words from the rest of the text.
• Infrequent capitalization: He rarely capitalizes the first-person pronoun “I” and often neglects to capitalize proper nouns, including names of people and places.
• Lowercase beginnings: McCarthy frequently starts sentences, including the first sentence of a chapter or section, with a lowercase letter.
• Emphasis through case: He occasionally uses all lowercase or all uppercase letters to convey emphasis or distinction, such as in the title of his novel “no country for old men”.
• Consistency: Despite these unconventional practices, McCarthy’s use of capital letters is consistent throughout his works, contributing to a distinctive and recognizable writing style.

What this tells me is that McCarthy uses capitals as a way to signify that some words are special. What that means may depend on the context. But I guess he isn’t using them just as sentence starters only, for example.

We can use the runic cross punctuation ᛭ when we want to signify that some text is special in some way, basically in the way that we would use captial letters. We wouldn’t use this to start all sentences, but maybe for names and certain terms (similar to the Shavian “namer dot” ·).

ᚦᛖᚾ᛭​ᛁᛚᚪᚱᛄ᛫​ᚸᚱᛇᛏᛟᛞ᛫​ᛗᛇ᛫​᛫​

Then Ilari greeted me.

ᛡᚣᚹᛟᚱ᛫​ᚠᚱᚩᛗ᛭​ᚫᛡᛟᚹᛟ᛬​ᛒᚢᛏ᛫​ᚫᛡᛗ᛫​ᚠᚱᚩᛗ᛭​ᛗᛁᛋᚳᛁᚸᛟᚾ᛫​᛫​

You’re from Iowa, but I’m from Michigan.

Now this isn’t to rule out that eventually we could have smaller case runes and larger case runes, similar to Cyrillic. But we certainly shouldn’t count on such a thing in the near term.

Question and Exclamation

Now what about question marks and exclamation points? Does McCarthy really avoid those too? I asked Meta to clarify yet again

• Rare use of question marks: McCarthy tends to avoid question marks, often replacing them with a period or a space, which can create a sense of ambiguity or uncertainty.
• No exclamation marks: He rarely, if ever, uses exclamation marks, which adds to the overall tone of restraint and understatement in his writing.
• Periods instead of question marks: When a sentence is structured as a question, McCarthy often uses a period instead of a question mark, which can make the sentence feel more like a statement or a declaration.
• Emphasis through sentence structure: McCarthy achieves emphasis through the structure of his sentences, using short, direct sentences to convey a sense of urgency or importance, rather than relying on exclamation marks.

Around the 700s, people used something like ᛫~,᛫˜, or ᛫̃, but that was replaced with the question mark that we see today by the 1600’s. So we could use ?᛫​ (or ᛫̃᛫) for a question. Basically the same as a full stop, but the first dot is replaced with a question mark.

But as mentioned above, how often is this even necessary? Most questions involve question words and phrasing, for example.

ᛞᛁᛞ᛫​ᛡᚣᚹ᛫​ᚸᚩᚹ?᛫​​​ᚻᚫᚹ᛫​ᚹᚩᛉ᛫​ᛁᛏ​?᛫​

Did you go? How was it?

If we absolutely need an exclamation mark, we could double up the comma. Or rather, double up the full stop ᛬​᛬, depending on how you look at it. Why not use ! since we could use ?. Well, ! supposedly has its origins in the latin letters “io”, so it’s a bit more latin-y in that sense.

ᚫᛡ᛫​ᛚᚢᚠ᛫​ᚦᚫᛏ᛬​᛬​ᛣᚫᚾ᛫​ᚫᛡ᛫​ᚻᚫᚠ᛫​ᛋᚢᛗ​?᛫​ᚦᚫᛝᛣ᛫​ᛡᚣᚹ᛫​᛫​

I love that! Can I have some? Thank you.

Ellipsis

Is there a pause longer than a full stop? An ellipsis … or em dash — could possibly fit that. Logically, this could be three space dots ᛫᛫᛫.

What would be a vertical pair for this one ⁝? Probably a semicolon, which is described as stronger than a comma, but weaker than a full stop.

ᛁᛏᛋ᛫​ᛋᚾᚩᚹᛁᛝ᛫​ᚫᛏ᛫​ᚦ᛫​ᛖᛖᚱᛈᚩᚱᛏ⁝​ᛣᚩᚾᛋᛟᛣᚹᛖᚾᛏᛚᛄ᛬​ᚩᚩᛚ᛫​ᚠᛚᚫᛡᛏᛋ᛫​ᚻᚫᚠ᛫​᛫​᛫​ᚫᛡ᛫​ᚹᚩᚹᚾᛏ᛫​ᛗᛖᛡᛣ᛫​ᛁᛏ᛫​ᚻᚩᚹᛗ᛫​᛫​

It’s snowing at the airport; consequently, all flights have… I won’t make it home.

Summary

In sum, what I am suggesting as the very basics is the following:

space ᛫
capital space ᛭
comma ᛬

full stop ᛫᛫

Cormac McCarthy could work with just that. I think it’s important to agree on some very minimal basics first, and let community usage determine the rest.

From here, people could add things. Stella’s system covers a lot of ground, for example.

Following the same logic as the above minimal suggestions however, here is an expanded set:

space ᛫
capital space ᛭
comma ᛬
semicolon ⁝

full stop ᛫᛫
question ​?᛫
exclamation ᛬᛬
interrobang ​?᛬

ellipsis ᛫᛫᛫

ᛒ​ᚫ‍ᛡ᛫ᛖᛗᛟᛚᛄ᛫​ᛞᛁᛣᛁᚾ‍ᛋᛟᚾ

Hope is the Thing with Feathers

by Emily Dickinson