Towards a Palantype Theory

[Jennifer Brien]

The Palantype is an interesting beast, an intermediary between Steno and Velotype. When it was invented in the 1940s the idea was that it should be possible to take dictation in a number of different languages and the paper tape that resulted should be readable with little training. 

Palantype dictionaries are all individual, there is no standardised theory as such. Here I will discuss the conventions used in the only dictionary I have access to – palantype-starter.rtf – so that you can follow them if they are helpful, or invent your own. 

Since raw Palantype is relatively easy to read compared with other machine shorthand systems, it is possible (and quite fun) to handwrite 'Palanese'. If you do this with your practice material, writing it as you hear it sounded, you will soon find what alterations you might consider when building your own dictionary.  

Difference in vowels between Steno and Palantype

Short vowels are normally has spelled if the sound is clear, but U is used for the indistinct 'schwa' sound as in the word 'a'. Because R comes later in the order: 

SCPTH+MFRNLYOEAUI^NLCMFRPT+SH

Palantype uses ^ to indicate a 'vowel R' if needed:  e.g. SCA^F for 'scarf',  CO^N for 'corn' and LU^N for 'learn'. U^ is used for all the indistinct R-flavoured vowel sounds like those in 'girl', 'surf' or 'fern'. The R key is used instead when the R is on its own or followed by letters later in the order. 

Long Vowels

Steno: AEU for the 'a' sound in words like 'grape' and 'saint'

Palantype: E^  (or EI for words that use two vowels, like 'seint') 

 

Steno: AOE for the 'ee' sound in words like 'seen' and 'preach' 

Palantype: I^  (or EA for words spelled that way)

 

Steno: AOEU for the 'i' sound in words like 'guise' and 'spite'

Palantype: AI

 

Steno: OE for the 'oh' sound in words like 'boat' and 'grown'

Palantype: OE (or OA for words spelled that way)

 

Steno: AOU  for two types of sounds:

The  'oo' sound in words like 'glue', but not when spelled "oo"

The 'ew' sound in words like 'few'

Palantype: OU for the  'oo' sound, and EU for the 'ew' sound

 

Diphthongs

Steno: AU for the 'aw' sound in words like 'bought', 'tawny', and 'faun'

Palantype: O^

 

 Steno: OU for the 'ow' sound in words like 'down' and 'mound'

Palantype: AU

 

Steno: OEU for the 'oi' sound in words like 'toil' and 'ploy'

Palantype: OI

Consonants

The extra keys make for a simpler mapping of sound to fingering. All the voiced consonants: B, D, G, V, and Z are produced by adding a + to their unvoiced counterparts P, T, C (pronounced K), F and S respectively. It's less usual in English to think of J as a voiced Y or NG as a voiced N, but those pairings make sense too, as does using Y to represent the ZH sound of 'Leisure' or 'Azure'. (Y is always a consonant).

That leaves a few consonant sounds that are represented by pairs of keys: 

Qu=CF  X=CS W=MF

 

 

All the left-hand combinations of consonant sounds are obvious and in order, with the exception of WH. (My accent is one of those that makes a distinction between the sounds of 'which' and 'witch', and so does Palantype). Some of these combinations are hard to key, though, so redundant keys may be dropped.  

HF stands for WH, SF for SW and TF for TW.

 

PM may be used for B instead of P+ on the left-hand side, and PH on the right.

The right-hand consonant combinations are also obvious, but there is no short equivalent of Plover's GS or BGS – instead the whole-stroke SH-N is used. Also, there is no 'j' sound on the right, so words like 'bulge' take an extra stroke P+UL/Y+.  I suppose you could argue that this sound is a voiced 'ch', which is available, but the general principle seems to be: when splitting words, always give the left hand a consonant if possible. 

There are some other things that can be done to simplify strokes:

Inversion of order is a lot less common than in Plover but it is used sometimes, as is the 'F for S' trick.

The voicing + may be dropped from the stroke is there is no risk of a conflict between the voiced and unvoiced consonant  e.g. 'baby' = PE^PI.

This seems to be another basic principle of Palantype: consider how the word may be stroked in full and then, if that poses any difficulty, add a shorter version dropping unneeded keys.

 

 

To illustrate this, here are the '100 common briefs' as defined in Palantype-starter. If you sound them out, you will hear that the accent if distinctly English:

'a':        U

'about':     A/P+AUT

'actually':    AC/TEU^/LI

'after':    AF/T-R    afternoon  AFT/N-N

'against':    A/CEN/ST

'almost':    OL/ST

'amount':    A/MANT

'another':    A/NU/TH-R

'any':    E/NI

'anything':    EN/THIN+

'are':            AR

'ask':            AS/-C

'be':            P+I^

'became':    PE/CE^M

'because':    COS

'become':    PE/C-M

'becomes':  PE/C-MS

'been':    P+I^N

'before':    PFOR

'beginning': PE/CNIN+

'being':    P+YIN+

'by':            PMAI

'can':    CAN

'continue':      CON/TNEU

'continued':   CON/TNEUT+

'could':     COUT+

'describe':    T+ES/CRAIP

'described'   T+ES/CRAIPT

'do':             T+OU

'easy':     EA/SI

'even':     I^/F-N

'ever':     EFR

'every':     EF/RI  everyday  EF/TEI

'example':     EC/S-M/PL

'have':     HAF

'having':     HA/FIN+

'he':             HI^

'idea':     AI/TEA

'ideas':     AI/TEAS

'if':             IF   if the IFTH

'in':             IN

'include':     IN/CLOUT+

'included':    IN/CLOU/T+-T+

'including':   IN/CLOU/T+IN+

'includes':     IN/CLOUT+S

'interest':     IN/TRE/ST

'I':             I

'is':             S

'many':     MNI   Minister MNI/ST-R

'move':     MOUF

'new':     NEU

'of':             OF   off O^F

'only':     OEN/LI

'opinion':     O/PI/NY-N

'other':     UTH

'others':     U/TH-RS

'otherwise':    UTH/MFAIS

'probably':     PRO/P+LI

'question':      CFE/STYUN

'receive':      RE/SI^F

'received':      RE/SI^FT

'said':      SET+

'somebody':     SUM/PTI

'somehow':      SUM/HAU

'someone': SM-N

'someplace': SUM/PLE^S

'something': SM-N+

'sometimes': SUM/TAIMS

'somewhere': SUM/HFER

'that':         THAT

'there':         THER there are the THRARTH

'therefore': THEFR

'the':                 TH

'they':         THEI

'this':                 THIS

'through':         THROU thoughout THRAUT

'to':                 TO

'today':         TO/TEI

'two':                 TOU^  too TOU

'understand':    UNS

'understanding':  UN+S

'understands':    UNT+/STANT+S

'unless':            UN/LES

'very':         FRI

'were':         MFUR

'what': HFAT

'where':   HFER     whereas HFE\RAS

'whether':   HFETH

'which': HFIH

'who': HOU

'with': MFITH

'within': MFI/THIN

'without': MFI/THAUT

'would': MFOUT

'yes' : YES

'you': EU

'you'd': EUT+

'you'll': EUL (you will)

'you're': EU^R your EUR

'you've': EUF

[Kell Willsen]

Thank you for this! I've been looking all over for Palantype resources, but could only find a reference to an out-of-print textbook from the 1940's.  

I am British, and sometimes have trouble finding the right vowel combinations in steno ("sayed" vs. "sed", "aey" vs "uh", etc.) 

I will continue to learn steno, and use my SOFT/HRUF with Plover, but your post makes Palantype a distinct possibility, rather than a distant dream. 

Thank you again.

[Jennifer Brien]

The Palantype starter dictionary available on Ted's Palantype download page at http://www.openstenoproject.org/palantype/assets/palantype-starter.rtf.  It is very British. 

RTF is a text markup format, so it can be read by any word or text processor. 

I've been wondering about the suitability of the HRUF as a Palantype. The + and ^ keys could perhaps be moved to the pinky and /or extended index of the top or third row, but how easy is it to press any of the central keys with a thumb, perhaps in conjunction with another thumb key?

If that isn't practical, then the Atreus looks quite promising. 

[Jennifer Brien]

Possibly the first effective real-time system used a Palantype machine and was devised back in the 1970's to assist deaf MP Jack Ashley. It didn't have a comprehensive dictionary, but used a small set of rules to transform the phonetic Palantype strokes into a readable pseudo-English like this:

THIS DEM STRAE SHN SHOES A PROE TOE TIPE AID FOR TH DEF WHIC HAS BEEN DE SINED AND BILT AT SOU THAMP TON UN VRESTI . TH DEFICE A TEMPTS TO PROE VIED A VI YOOL TRANS CRIPT OF LEC TURES ETS . FOR TH DEF UN FOR TNAT LI , NOR MAL SPEECH IS TOE RAPID FOR A TIE PIST . THEFR TH OE PRAE TR OF THIS MA SHEEN IS A SHORT HAND RIE TR . SHEE IS UE SING A PA LAN TIPE MA SHEEN TO TACE DOUN THIS LEC TURE


This demonstration shows a prototype aid for the deaf which has been designed and built at Southampton University. The device attempts to provide a visual transcript of lectures etc. For the deaf unfortunately, normal speech is too rapid for a typist. Therefore the operator of this machine is a shorthand writer. She is using a Palantype machine to take down this lecture.

from Design and the Digital Divide: Insights from 40 Years in Computer Support ...
By Alan F. Newell

Not bad for 40-year-old tech! More or less all that was done was to decode punctuation and characters represented by two keys. There were at the time less than 100 trained Planatypists in the UK and the machines were no longer being produced, so no attempt was made to vary the keying for a more orthographic output and although the software has made some rudimentary attempts at combining syllable into longer words, the style of 'maximum onset' (keying as many consonants as possible with the left hand even at the expense of extra strokes) is clear.

This got me thinking that the Palantype layout might be a useful basis for a orthographic system, so you could type anything 'as written' but resort to a full dictionary when needed. The aim is to have it as easy to use as a normal keyboard for normal typing, using minimal cues such as keycaps and cheat cards until speed and confidence is built. I'm now convinced that more keys are better – up to a certain point. The 48 keys of a full Planck Grid are probably too many. Planatype uses 29 logical keys, so there is some room for expansion, but each extra key has to earn its place.

A few more consonants are needed – notably K and right-hand W.
It's obvious to have W as MH on both sides, and PF looks the only possibility for K on the left (by analogy with HF for WH etc. we could then use FN for KN). On the right the combination FH is free for K, and slightly easier to chord than FP, as well as distinguishing between MFP for WP as in Gawp, and MFH for WK as in Hawk. On orthographic order, right hand R and W precede all other letters, allowing RN, RL, RC, RM, RF, WN and WL.

+H looks useful for the common right hand GH combination, +TH, being GHT.

We definitely need a NOSPACE key to mark where syllables are joined. Simple words, where all the keys are in steno order can be output directly after doing the sort of character substitution shown above. If the stroke does not include NOSPACE, then a space is automatically added. Obviously. on a Qwerty keyboard this will be the spacebar (on its own, NOSPACE ouputs a space). If a chord does not include NOSPACE or directly follow a chord that contains NOSPACE, then the dictionary is used, otherwise it is output according to orthographic rules.

NOSPACE can be combined with punctuation to reduce the number of chords that need to be learned. For example, ^ on tis own in Planatype is a period (space and capitalise next) NOSPACE^ would be a dot separator.

Most punctuation would be strokes on their own (I've tried merging . and , but it doesn't seem worth it). The only reasonable exception is apostrophe, which seemed to belong after the vowels in the steno order, so that many common contractions can be written in a single stroke.

A dedicated Index Case Key would be useful to, to Capitalise the first letter of the stroke it is combined with, to CAPITALISE the entire following stroke (and any linked to it by NOSPACE if it is used on its own, and to otherwise do what Shift does on a normal keyboard.

Even with just that I think it would make a usable orthographic system, but a few more dedicated keys might be added from Jackdaw:

E to follow right hand consonants
Y to follow right hand consonants, that becomes an I with NOSPACE
* to double single vowels, reverse the order of two vowels, and perhaps code the few three-vowel sequences that are not in steno order.
+ to double consonants following short vowels

Since these last two are meaningless without vowels (as are Palantype's + keys with only consonants of the 'wrong' side) they can be as 'layer' keys to, for example, bring the number pad cluster under the right hand.

Dedicated keys for the inflections of plural S, -ING, -ED and possibly -ER are more luxuries. Common postfixes can be handled either orthographically or by using the dictionary.

[Gabriel Holmes]

How about a foot pedal for NO SPACE?

[Jennifer Brien]

If you're building a keyboard then palm switches such as Velotype uses for Cap and No Space seem like a good idea, but I was thinking of something that could be done in software, as an extension of Plover.

I /think/ any orthographic theory can be defined by combining the output from several micro-dictionaries, but I haven't quite worked out the syntax yet.