Towards an English Orthography

[Jennifer Brien]

An orthographic theory of shorthand is based on spelling rather than phonetics. The advantage is that you can (in theory) write any word you know how to spell at least as fast as you can type it, because you do not need a dictionary to convert it from phonetic to written form. It should be easier to learn, and quicker to get to a speed that is usefully faster than qwerty.

This is a sketch for an orthographic theory that might perhaps be implemented on a HRUF or a Planck the like.

There are three lookup tables: one for the left hand, one for the thumbs, and one for the right hand. That's all, apart from a few extra control keys and some logic to combine them.

Vowels

With only four keys to represent the vowels, only the ten most common vowel pairs can be produced in one stroke, because here are eleven possible combinations of more than one key and we need one of those for the missing single vowel. Fortunately the top ten vowel pairs account for nearly 90% of all vowel pairs in typical text, so perhaps that is good enough. A little logic will provide some more.

Following Plover tradition, they can be coded thus:

  AI=AEU   AU=AU

  EA=AE    EE=AOE 

  I=EU     IE=OE    IO=AO

  OI=OEU   OO=AOU   OU=OU

Opening and Closing Consonants

The standard key assignments are well chosen for phonetics, but not not so good for orthography.

There is an orthographic layout, obviously inspired by stenotype, in a patent for a 'syllabic typewriter' filed by Dothan Shelton in 1976. http://www.google.com/patents/US3970185

The keys are: 

Left Hand    Right Hand

     ACWN    RLCTE

     STHR    NGHSY

Note the extra A on the left hand, and the E on the right. A is the most common initial letter apart from T, and it makes it possible to stroke such common words as AGAIN, ABOVE and ALONG in one. Quite a few common consonant sequences are immediately obvious.

Here is a spreadsheet of the full left and right-hand tables:

https://docs.google.com/spreadsheets/d/1-IvW_llymk3bNv_L2tILHWODE8RSaGDoGorNbeO3Otk/edit?usp=sharing

I've put the middle finger letter on the y axis and the outer fingers on the x axis, because I think that best shows the 'family resemblances'  The less obvious combinations are highlighted in yellow. It's not necessary to learn them all at once - you could possibly just learn the single character chords and still be able to type at a good speed.

The missing consonants are:

Left hand

  B=CTWH D=CT  F=CTH G=SCT

  J=TWN  K=TWH L=NR  M=WN

  P=CW   Q=TWN V=TN  X=STW Y=HN Z=CN

Right hand

   B=GC D=NLG or CHS   F=GCH K=GT M=NGH

   P=LC S=NL or S      V=NH X=LGH Z=LN

Generally, what is typed is the output of the left-hand chord (if any, followed immediately by the output of the thumbs (likewise) and the right hand (ditto).

Control Keys

Since we want to do more than just type characters, it is useful to have extra keys, like steno's * key, that can be part of any stroke. It is possible to have six more of these fairly easily: a double-height key each side of the left hand, and a double-width key above and below each thumb position. Up to four of these can be in use simultaneously no matter what the stroke (obviously you can't operate both control keys for  one thumb at the same without getting a vowel too). in some cases it might be useful to have a control key that may, or may not be part of a stroke (think the number keys on a steno keyboard)

What do they do?

Caps

If the Caps key is repressed and released between the end of the previous stroke and the end of the current one, uppercase the first letter in the output.  If it is held down throughout, uppercase all the letters.

Punctuation

' is very common in informal English, and might usefully be a control key.  Used as part of a chord it inserts itself as the second-last character, unless it recognises a two-letter contraction. On its own it's just another character.  Really tricky things like fo'c's'le would be four strokes, divided as fo'c|'s|'|le rather than the seven needed if each apostrophe was stroked separately. 

. and , should be on control keys if possible, since they are the most commonly used punctuation. They are in fact (like some other punctuation) special cases of SPACE. 

SPACE 

As part of a chord, precedes it with a space. Velotype takes the view that a well-formed syllable (a chord containing a vowel) is much more likely than not to be a word, so a space is automatically added in all those cases, and SPACE with a syllable removes the space to form multisyllabic words. If I were doing that, I'd be inclined to have only the thumb keys count as vowels for that purpose. Experiment will decide.

*

The Swiss Army knife. On its own, as in Plover, it deletes the last chord. In conjunction with thumb keys, it doubles the following consonant. With finger keys, it checks the output of your mousing hand and uses that to switch the table for the non-mousing hand, allowing access to specialised keys and chords. The table stays in force util either the * key is released or the mousing hand selects a different table.  If there are only keys on the non-mousing side then they are interpreted as mirror images of the other hand, and the new table is active on the next stroke. The idea is to use this for numerics, function keys, cursor movement and the like. You could also use it for briefs - why not?

Suffixes

The common suffixes -ED  -ING and -LY might be put on control keys (the right-hand table handles plurals quite well already) in that case -LY should implement the 'c+ly -> cally' rule.

Letter-combining rules.

Whenever you can't get all you want in one stroke, break it in two. For example, the most common missing vowel pair is OA, so you would have to stroke COAT as CO|AT.  Sometimes it helps to use the end vowels; most uses of UE and OE (where they are not in two separate syllables) are at the end of words, so the trailing E can be used.  Q always adds a U if followed by another vowel (but what about Qatar?)

Trailing Y, if preceded by a vowel and at least one consonant changes to I if it is not at the end of the word. Thus, BUSINESS is stroked as BUSY|NESS. if you really want a Y there, use the left hand Y instead: BUS|YNESS.

That's all I can think of for now.

Comments?

[Hanchul Park]

I am impressed in two things. The first is, that this system does not necessarily need a dictionary as far as I understand, like Kingslet.

The second is that it involves more keys. When I saw steno the first time, it looked weird since it is not symmetric. I know more keys don't mean better, but our left hands can do more than now. But I am not sure if the control keys for thumbs can be pressed conveniently or not.

2016년 1월 17일 일요일 오전 4시 26분 37초 UTC+9, Jennifer Brien 님의 말:

[Zack Brown]

Hi Jennifer,

This looks fascinating! Sort of a cross between Kinglet and Plover.
You've kept Plover's syllabic form and Kinglet's predictable
orthographic chords, but taken away Plover's extreme memorization
requirements and Kinglet's extreme dexterity requirements. It seems
like exactly what you were looking for in your initial responses to
Kinglet.

Just from glancing over your materials, it seems like you must be
right about getting a significant speed increase over QWERTY.

In the best case, it looks like you can get 7 or 8 characters per
stroke or more, which probably adds up to being even faster than
Plover for ordinary English.

The system may run into problems with less common character sequences.
But if you can overcome those, I'll bet it would be blazingly fast,
and still relatively easy to learn.

Be well,
Zack

> --
> You received this message because you are subscribed to the Google Groups
> "Plover" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to ploversteno...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.



--
Zack Brown

[anobo]

Hi Jennifer,

wow that sounds great.

Let me confirm:

- the order of the letters is still like in steno, i. e. A S C T ... 

- inversion would be rather avoided (or even be not allowed?)

With regard to the control keys, these could be also used for special characters in other languages, such as the dots above a, e, o, u (e. g. German Umlauts).  Admitted, the letters itself may need some further arrangement depending on the language.

In general this raises the question whether anybody knows some sources/statistics with regard to this:

Which letter (combination) is used how often on which position (and that by language)?

[Jennifer Brien]

On Sunday, 17 January 2016 09:21:21 UTC, anobo wrote:
> Hi Jennifer,
>
>
> wow that sounds great.
>
>
> Let me confirm:
> - the order of the letters is still like in steno, i. e. A S C T ... 
> - inversion would be rather avoided (or even be not allowed?)
>

No, there is no 'steno order' as such, though natural left-to-right order is preserved where possible. There is no diction, so inversion is not possible. The keys simply act as an index to the appropriate table, as in the spreadsheet I linked to. Any combination will always return the same result, unless the table is changed.

>
> With regard to the control keys, these could be also used for special characters in other languages, such as the dots above a, e, o, u (e. g. German Umlauts).  Admitted, the letters itself may need some further arrangement depending on the language.
>

Well, yes - but let me deal with English first! The basic premise is that there are a limited number of common openings to a syllable in any language, a limited number of common vowel sequences,and a limited number of common closings. Put these together and you should be able to type most monosyllabic words in one stroke. The control keys are for non-alphabetic characters that can appear /at a known position/ in relation to any alphabetical string (or, in the case of a Caps key, modify it in some known way) It's useful if it can be keyed at the same time as the syllable to save a stroke, but there are a limited of positions for a key that can be pressed in conjunction with the keys for /any/ syllable.

I've tested this against a few short passages of Steno,and it seems that it can beat Plover for stroke count if Caps, Space, ., ,, ', and -ED are all control keys, but the strokes are probably more fingery, so I wouldn't expect it to be ultimately as fast. That's not the intention. This is more a rival to Dvorak and other ergonomic character-by-character layouts - "Steno for the rest of us"! ;-)

[Gabriel Holmes]

I have somewhat mixed feelings about orthographic approaches (even compromises) because:

1. If your objective is to keep up with human speech, then you want to go directly in your brain from sound to stroke, without wasting time and brain energy on the spelling.
2. Orthographic systems seem to occur more for languages that are far more phonetic than English.

[Zack Brown]

Hi Gabriel,

I think your item #1 is not accurate. The vast majority of people using a chorded typing system in the future, will not be transcribing human speech. They'll be using it the way people use QWERTY keyboards now - to write things. Emails, chats, novels, essays, computer programs, and so on.

So the need for pure speed just isn't there. As a professional writer myself, I can say that I just want enough speed to keep up with my own thoughts. That's probably the situation of the vast majority of people considering a chorded typing system. Plover, Kinglet, Jennifer's system, and other free chorded typing systems, are going to be reaching out to a far broader audience than the steno systems of the past. Now that people don't have to pay $25,000 to get up and running, there doesn't have to be a job associated with it. People can use it just because it's better.

Meanwhile, the learning curve for Plover, even with all the tutorials and practice aids out there, remains extremely steep. In order to get up to a decent speed, you have to memorize a huge number of chords that go beyond the fundamental phonetic theory. It takes a tremendous commitment to get over some of those hurdles.

The thing you'll notice about orthographic systems like Kinglet and Jennifer's system (we could call it JenSys if she doesn't come up with another bird name), is that there is very little to memorize. A few dozen or a hundred chords. Compared to Plover, the orthographic systems have no learning curve at all. It's that big a difference.

Don't get me wrong - Plover's great! Whoever finally cracks 400 WPM will probably be using Plover. And clearly there are plenty of people who like the phonetic/mnemonic approach. It's a very cool system with a lot of good stuff in it.

But it seems like you're saying there's no value to having an orthographic system. And clearly there is. For speed approaching Plover's, for ease of learning, for the ability to type whatever you want instead of just plain words, and for the ability to think about words in the same way we our teachers taught us to think about them (i.e. as things that are spelled), an orthographic system is able to address all of those needs.

Be well,

Zack

--

You received this message because you are subscribed to the Google Groups "Plover" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ploversteno...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Zack Brown

[Theodore Morin]

The main difference between steno and orthographic systems is that steno has existed for a long time and is in wide use. The exception is the Velotype, which I don't know enough about to comment on in a productive way.

I'm really excited for people to hit 30WPM on these new and creative orthographic systems. Then, I feel, we'll be better able to compare the two.

The learning curve for Plover is only theoretically great. I found that in practice I didn't ever have to consciously write a list of chords down and memorize them. Only very rarely did I get confused with a set of chords, like the set for "identify" as a root, but 5 minutes of learning and practicing got me over that. Plover starts very slow, but I'm not certain that it is slower to learn overall than an orthographic system.

The only other factor that comes to mind is whether Kinglet et al would benefit from specialized hardware, which steno definitely does.

[Jennifer Brien]

On Friday, 22 January 2016 14:48:33 UTC, Zachary Brown wrote:

Hi Gabriel,

The thing you'll notice about orthographic systems like Kinglet and Jennifer's system (we could call it JenSys if she doesn't come up with another bird name), is that there is very little to memorize. A few dozen or a hundred chords. Compared to Plover, the orthographic systems have no learning curve at all. It's that big a difference.

I think I'll call my system Jackdaw, because I think it's clever, and it steals a lot of shiny stuff from other sources. :-)

The basic goal is that it should be at least as fast as Qwerty for all input, and faster than the best ergonomic layouts for anything remotely pronounceable. Take for example the Welsh village with the outlandish name:

   Llanfairpwllgwyngyllgogerychwyrndrobwllllantysiliogogogoch

That's awkward because w is a vowel in Welsh (roughly double-u) and there's no way to have a double L on the left hand, but the rule is 'do what you can with each stroke', and it doesn't matter if the consonants are from the left hand or the right (apart from the rule about trailing Y changing to i in the middle of a word) so it can be divided as:

   Ll|an|fairp|wllg|w|yng|yll|goge|r|ych|w|yrn|drob|wll|ll|ant|y|sili|og|og|og|och

22 strokes for 58 characters. Not bad. It could be three less,, because left-hand WY and RY look possible, but Shelton table has them produce MY and LY respectively. That can be changed - it all depends on whether you are more likely to write about wyverns than myopia or Ryan than lying.  If you write a lot of Welsh names, you might want to assign a lefthand chord to produce LL, and that would save another two strokes.

You see what I mean about it being like a jackdaw?

[Gabriel Holmes]

Hi Zack,

Reading back on what I wrote I'm a bit ashamed. I absolutely did not mean to insult orthographic systems, just that I'm a bit skeptical of them. Also, I think the line between phonetic/orthographic is rather blurry anyway, don't you agree?

G

[Zack Brown]

Sorry if I came off too aggressive about it. I've been wanting an orthographic system for so many years, I've got all my reasons down by rote. But you're right - there's a lot to be skeptical of as well. In Kinglet's case, maybe the chords can never be made simple enough. In Jackdaw's case, maybe the speed won't get enough beyond QWERTY. And in all cases, maybe Plover is just better anyway.

If nothing else, I think it's great that the Plover software is evolving to support these various new systems. I'd imagine there'll be tons more coming out in the next few years. Plover dictionary files for various languages; new orthographic systems; and who knows what else. The main thing is just that it'll probably be a lot of fun for a lot of people. And whatever ultimately emerges will be a surprise for everyone.

Be well,

Zack

[Zack Brown]

Jackdaw! A great name! Let me know if you'd like me to document it on the Learn Plover pages, next to Kinglet.

(the current incarnation of Kinglet can do 'Llanfairpwllgwyngyllgogerychwyrndrobwllllantysiliogogogoch' in 12 strokes, but that may be changing soon, as I search for more typeable chords)

Be well,

Zack

--

You received this message because you are subscribed to the Google Groups "Plover" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ploversteno...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Zack Brown

[Zack Brown]

Hi Theodore,

For Kinglet, specialized hardware would definitely help. The chords are so difficult in the current incarnation, that regular QWERTY keyboards make it all the more error-prone. Even with easier chords, Kinglet will still need a keyboard that makes it easier for a single finger to press neighboring keys. I've been thinking of designing my own keyboard, maybe using some of the wiring ideas from existing Plover keyboards.

Does Velotype have any really solid documentation available? I haven't been able to find anything that really explains the theory.

Be well,

Zack

[Charles Shattuck]

I looked up Velotype the recently and it sounds pretty good. Is there some reason that system can't be used? Is it patented?

[Achim Siebert]

So true - first of all it's lots of fun. But I'd never have started learning Steno if it went by English orthography. If it were Spanish, French, even German - yes, it might make sense. But I'm glad for English it's WYHIWYG (what you hear is what you get), since English orthography is completely unreliable, since some words are of Latin origin, some French, some Germanic, and thus it's simply too chaotic orthographically. If I know an English word only by its written form, I sometimes even learn how to pronounce it correctly by looking at the stroke definition!

[Theodore Morin]

Even as a native English speaker, steno has taught me how to pronounce some words :)

[Hanchul Park]

As a foreigner once imagined studying English by studying Plover, this is great!

2016년 1월 24일 일요일 오전 9시 4분 48초 UTC+9, Ted Morin 님의 말:

[Gabriel Holmes]

As I've said before, hopefully steno will eventually be taught in elementary school and first graders will come up with far better theories than any of us old farts ever could.