"Sagispeak"

[cam.taylor]

We are hoping to have fixed spelling across all languages that use latin/roman characters, and in order to do so we are happy to allow different pronunciations

I'm assuming this also applies to those "+h" mixed phonemes? I think it's a really cool idea with those hs, I was seriously wondering why "ph" and not "f" was a thing, but that's explained it to me. But thinking about those phonemes, ph, sh, kh, and even th (for 49:48, 96:49, etc), their pronunciation will differ across languages, as you've said. That's cool. But some languages don't have defined pronunciations for these special combinations, and so they might have to share pronunciations with simpler phonemes.

Just thinking out loud here, and I'm sure you've probably gone through a couple of these, but here goes.
How would someone who's never studied a non-English language pronounce "kh"? /kx/? /kəh/?
These combinations make a WHOLE lot of sense in Korean (though very different to the English morpheme logic), but hardly any in Japanese, where ph would generally be p(u)f(u), and th would be ts(u)f(u), but when adding direction p(u)hai/p(u)hao and ts(u)hai/tu(u)hao. I'm assuming we want to keep this mostly in roman script to make sure nothing gets too mangled in "translation".

From my very limited knowledge of French, I think the "j" phoneme you alluded to is /ʒ/, as in English "vision" [vɪʒən]

That's one I championed for some time too. But George convinced me that it was more important to preserve the alternative pronunciation for "j", and the digraph spelling of the double-same-flags. And I think an initial "h" is silent in some languages. It is silent in some dialects of English. So George convinced me to accept "n" for , which is like the symbol with (most of) it's shaft missing, as are some other short-ASCII characters for Sagittals.

Agreed! This all seems very well thought out, as I'm sure it's taken a lot of deliberation.

If we have vai and vao for and respectively I thought it would be quite elegant to use wai and wao for and . It's a symbol that gets used moderately often, although perhaps and would be prioritised first for single syllable short-names. That is, IF there are excess general phonemes left over after naming all of the Spartan set flags.
I know it makes sense to call them by the sum of flags, but part of me is quite averse to calling "jakai", etc. Then again, maybe I'll live with it. Priorities first.

I'm off to bed now but I have a bit more to mull over. I'll try and post again tomorrow. I'm quite liking the consonants you've already assigned, although having one voiced/unvoiced pair (z/s) might be hard for some languages that don't distinguish. Like Maori. But even then, modern speakers seem to be using more and more loanwords using the English, not Maori approximate, pronunciation. I really liked how s and z are a pair that sound similar but have the opposite feature of voice, and the symbols look similar but are oppositely left/right. Really nice.

Hopefully I'll get back here tomorrow. Cheers.

[Dave Keenan]

But thinking about those phonemes, ph, sh, kh, and even th (for 49:48, 96:49, etc), their pronunciation will differ across languages, as you've said. That's cool. But some languages don't have defined pronunciations for these special combinations, and so they might have to share pronunciations with simpler phonemes.

Just thinking out loud here, and I'm sure you've probably gone through a couple of these, but here goes.
How would someone who's never studied a non-English language pronounce "kh"? /kx/? /kəh/?
These combinations make a WHOLE lot of sense in Korean (though very different to the English morpheme logic), but hardly any in Japanese, where ph would generally be p(u)f(u), and th would be ts(u)f(u), but when adding direction p(u)hai/p(u)hao and ts(u)hai/tu(u)hao. I'm assuming we want to keep this mostly in roman script to make sure nothing gets too mangled in "translation".

There is no sagittal !)) (which would be "thao"). For various reasons, the 49-medium-diesis symbols is which is "jpao" /jəpaʊ̯/.

When possible, I think "ph" should be pronounced /f/ or /ɸ/ and could be respelled "pf" in German. "sh" should be /ʃ/ and could be respelled "sch" in German. "kh" could be pronounced either /x/ or /tʃ/ and would need to be respelled "ch" in English and German, and "tch" in French, to ensure this. Languages which don't have these phonemes can pronounce them /pəh/ /səh/ /kəh/.

From my very limited knowledge of French, I think the "j" phoneme you alluded to is /ʒ/, as in English "vision" [vɪʒən]

Yes, that's it. Thanks.

If we have vai and vao for and respectively I thought it would be quite elegant to use wai and wao for and . It's a symbol that gets used moderately often, although perhaps and would be prioritised first for single syllable short-names. That is, IF there are excess general phonemes left over after naming all of the Spartan set flags.
I know it makes sense to call them by the sum of flags, but part of me is quite averse to calling "jakai", etc. Then again, maybe I'll live with it. Priorities first.

Unfortunately we need consonants not only for the flags in the Spartan set, but for the other 3 flags as well: . But I agree it's a shame if that means we have to do without single-consonant pronunciations for the multi-flag Spartans. There are problems because various languages pronounce "w" as /v/, "v" as /f/ and "z" as /s/ or /dʒ/. We may have to lose the option of as /v/.

Please don't hesitate to criticise even our existing choices.

[cam.taylor]

Dave, I've finally seen the full updated "smilie" list, with names. Looks amazing, and answers a LOT of questions. I started writing a post about the pronunciation of unusual flags and diacritic marks for usage in discussing very high precision intonation, but found my question mostly answered by hitting the "view more smilies" button and looking at page two. Impressive.

Is there any way to put all the smilies on one page, so we can see them along with pronunciation all without hitting pop-up boxes? Maybe this web format makes things difficult, but that sure would be nice.


I see you've got the slash diacritic pronounced as a medial "r" or "l". Quite like that. This is incredibly high-precision territory and the "pronunciation key" packs a WHOLE lot of information into that little word.

I also see you've got as "o" (the default low vowel) and as "i" (the default high vowel). Very very good, but how do they connect with regular flags' pronunciation? Before? (After certainly wouldn't work) Is E (a pythagorean diminished fourth from C spelled as a schimatic major third) "E opao"? Is a pythagorean comma above C then C "C ipai"?...
What happens when we have two accent marks, does that make a double-vowel? Long vowel? as in "E oopao" or "E ōpao"? Sorry if this is getting a bit too detailed, I'm just getting excited and trying to learn the whole system. It looks very logical and pretty intuitive.


Just one little thing I have when combining a flag and an apotome symbol with a nominal or interval size. I think they should be combined in that order, i.e. flag name *before* flat/sharp/double flat, or interval name (e.g. minor 6th) etc. So "E pai flat" and "pai minor third" (or possibly "pai flat third") rather than "E flat pai" and "minor third pai"? (flat third pai?), this second order sounds very weird to me. This is also the order they're written on manuscript using mixed Sagittal, although of course the notehead there comes LAST, as in "pai flat E", which is a little counterintuitive to say.

[Dave Keenan]

Is there any way to put all the smilies on one page, so we can see them along with pronunciation all without hitting pop-up boxes? Maybe this web format makes things difficult, but that sure would be nice.

We're preparing a post for this thread that will do that, and will also give the upward ASCII shorthand characters for the spartans and athenians. You can only get the downward ones from the sagispeak names. And we'll make it accessible directly, via a link from the main Sagittal web page.

It has been delayed by something you wrote in your interval naming post which I thought was obviously a good idea and am waiting for George to approve. You should know that the optional wai wao for is purely the result of you suggesting it. We were holding "w" in reserve as an alternative for in languages that don't have "z", or don't distinguish its sound from that of "s" (Spanish), or "j" (I forget which langauge(s)). But George realised we could just recommend respelling "z" as "ts" in those cases. "ts" doesn't occur otherwise, because the "s" and "t" flags are on opposite sides, and in the opposite order.

I see you've got the slash diacritic pronounced as a medial "r" or "l". Quite like that. This is incredibly high-precision territory and the "pronunciation key" packs a WHOLE lot of information into that little word.

That's not actually a diacritic. It's a flag -- the left scroll (left concave flag) or 19-schisma flag. It looks like a diacritic when it appears in conjunction with a left barb (left straight flag). However the true diacritics never actually touch the core symbol or shaft that they appear with. But yes, it's brilliant how "r/l" blends with so many other consonants, because left-scroll is combined with so many other flags in the high-precision Promethean extension, including the few 3-flag symbols.

We're happy for "r" to be pronounced in any of the many ways it is pronounced in various languages. It is even OK for it to stray over into "kh" territory, as it does in French, where it is almost pronounced /x/, because in French "kh" should be respelled as "tch" anyway. And these consonants are only used for very-rarely-used symbols.

I also see you've got as "o" (the default low vowel) and as "i" (the default high vowel). Very very good, but how do they connect with regular flags' pronunciation? Before? (After certainly wouldn't work) Is E (a pythagorean diminished fourth from C spelled as a schismatic major third) "E opao"? Is a pythagorean comma above C then C "C ipai"?...

You got it.

What happens when we have two accent marks, does that make a double-vowel? Long vowel? as in "E oopao" or "E ōpao"? Sorry if this is getting a bit too detailed, I'm just getting excited and trying to learn the whole system. It looks very logical and pretty intuitive.

It's very pleasing that you are so excited. And even more pleasing that it's intuitive enough that you can just guess how it works. But in this case I must disappoint you, because there are no symbols with double 5-schisma diacritics (about 2 cents each). However, if there were, I would totally go with what you've just proposed, with the long vowels.

However, up in the realm of the gods, in the Olympian (extreme precision) notation, there can be symbols with one or two 13-limit schismina diacritics (called "mina" diacritics for short, and worth about 0.4 cents each). At present these are represented by using the same diacritics, but putting them to the right of the core symbol (or bare shaft). And so, their syllable comes at the end of the word, even though it's possible that in future we will instead use a different shaped diacritic to the left of both the core and any schisma diacritic. But as you noted, we can't just add a bare "o" or "i" to the end of the word. George's solution was to insert a consonant that we haven't used for any flag, or for any pair of flags in spartan. One consonant for a single mina diacritic and another for a double mina diacritic. This is the least used part of Sagittal, so it got the consonants that were left when everything else had been assigned. We used "m" for one, and "b" for two. Fortunately those have some obvious mnemonics: mono and bi, or mina and bi-mina.

Just one little thing I have when combining a flag and an apotome symbol with a nominal or interval size. I think they should be combined in that order, i.e. flag name *before* flat/sharp/double flat, or interval name (e.g. minor 6th) etc. So "E pai flat" and "pai minor third" (or possibly "pai flat third") rather than "E flat pai" and "minor third pai"? (flat third pai?), this second order sounds very weird to me. This is also the order they're written on manuscript using mixed Sagittal, although of course the notehead there comes LAST, as in "pai flat E", which is a little counterintuitive to say.

First, I must point out that the term "flag" is used to refer only to those parts of a Sagittal symbol that are not shafts or diacritics. There are 8 types of flag, corresponding to what would be left if you removed the shafts (the long vertical lines) from these 8 symbols:

So symbol = shafts + flags + diacritics.

I totally agree that in the case of the interval names, the sagispeak word should come first. But in the case of the pitch names, George and I agree (at least in the case of mixed Sagittal) that the sagispeak word should come last. We reason by the following analogy. As you point out, although we write it on the manuscript as O where the "O" represents an E notehead, we pronounce it as "E flat", i.e. the least significant part is said last. So what is notated on the staff as O should be pronounced "E flat pai". Another reason is that, to many musicians, "E flat" is just another note name in its own right, not really considered as a modification of an E. And so "E flat pai" is read as "(E flat) pai" and therefore it would be completely unnatural to insert the word "pai" between the "E" and the word "flat".

In the case of the pure Sagittal, i.e. using multi-shaft sagittal symbols and no conventional sharp or flat symbols, George thinks the multi-shaft symbols should be pronounced the same as their mixed counterpart so that O is "E sharp pao", same as O. But I think it would be good to distinguish the multishaft symbols by reversing the order and making a single word, so that O is "E paosharp".

Of course you are free to use them however you see fit.

[George Secor]

Sagittal Shorthand and Sagispeak
--------------------------------
19 Sep 2015
 
Following is a plan for single-character ASCII shorthand for single-flag and athenian-level single-shaft symbols, integrated with spoken names for all single-shaft Sagittal symbols.
 
Sagispeak symbols not containing a mina diacritic (a.k.a. right-accent) always end with a diphthong, either "ao" for downward alteration (as in "dOWn") or "ai" for upward alteration (as in hIGH). The diphthong occurs once and only once in a symbol name, and it indicates the overall direction of pitch alteration. The consonant(s) preceding the diphthong are determined by the downward shorthand characters of the constituent single-flag symbols according to the table given below. An exception is that the shorthand character "\" (for 5-comma down) becomes the consonant "p" ("pental"). Thus is named "pao", and is "pai".
 
The number of consonants in a symbol name will generally be the same as the number of flags in the symbol and will occur in the same order (left-to-right).  For a symbol having two flags on opposite sides of the arrow shaft, the consonants in the name are separated by the vowel "a" (pronounced "AH").  Thus, since (7-comma down) uses shorthand character "t", is named "patao", and is "patai".

If a symbol has two flags on the same side of the arrow shaft, then the two consonants are not separated and are blended where possible, as shown in the following table.  If a symbol contains a double flag (same flag type on same side), then the corresponding single-flag consonant is combined with "h" to form a double-letter consonant.  Thus is named "phao", and is "phai".
 
If a symbol contains a schisma diacritic (a.k.a. accent), then the prefix "bo" (down) or "bi" (up) is added to the symbol name. Thus (diaschisma down) is named "bipao", (pythagorean comma down) is "bopao", and (diaschisma up) is "bopai". If a symbol contains no flags, i.e., only diacritic(s), then the last diacritic vowel "o" or "i" becomes "ao" or "ai" to satisfy the condition that a diphthong must occur once in a symbol name. Thus is named "bao", is "mao", and is "mibai", but (11:17M diesis, 33:34) is "mibopakai".

Plain-text
shorthand      Sagispeak consonant(s)
---------------------------------------
Spartan symbols:
n  u      	nao nai                       [Edit: In the original post, "c" was given instead of "u"]
\  /      	pao pai
t  f      	tao tai
_  =     	phao phai (pronounced fao fai) (pp -> ph)
&  %     	patao patai or gao gai
v  ^     	pakao pakai or vao vai
w  m     	jatao jatai or wao wai
d  q     	jakao jakai or dao dai
Non-spartan single-flag symbols:
r  ;      	rao rai (may be pronounced lao lai)
z  ~      	zao zai (may be pronounced tsao tsai)
k  y      	kao kai
j  ?      	jao jai (j may be pronounced as in jaw, yaw, haw, or zhaw)
s  $      	sao sai
Remaining athenian symbols:
i  *     	ranao ranai
o  e     	sanao sanai                     [Edit: In the original post, "u" was given instead of "o"]
a  g     	janao janai
Non-athenian blended symbols:
          	shao shai (schao schai in German) (ss -> sh)
         	chao chai (tchao tchai in French) (pronounced tsh) (kk -> kh -> ch)
          	slao slai (sr -> sl)
          	prao prai
         	phrao phrai
         	prakao prakai or vrao vrai
Remaining (non-athenian) compound symbols:
         	razao razai
         	ratao ratai
         	satao satai
         	pazao pazai
         	sakao sakai
         	jazao jazai
         	jpao jpai (pronounced jəpao jəpai)
         	ktao ktai (pronounced kətao kətai)
         	rachao rachai
Schisma diacritics:
.  '      	add "bo" or "bi" prefix ("bi-cent")     [Edit: The original had o/i instead of bo/bi]
Schismina diacritics:
,  `      	add "mo" or "mi" prefix (1 mina) 
,, ``    	add "momo" or "mimi" prefix (2 minas)   [Edit: The original had bo/bi instead of momo/mimi]
Apotome accidentals:
b  #     	flat, sharp (or equivalent terms in other languages)
bb x     	double flat, double sharp
h        	natural

Multi-shaft (or "pure") Sagittal symbols have the same names as their mixed-symbol counterparts. Thus F and F are both named "F-sharp-pao" in English and may be written "F#-pao". Since most other languages do not use the terms "sharp" and "flat", the Sagittal name may be appended to the appropriate pitch names, separated by a hyphen.

[cam.taylor]

Dave, Wow, there's quite a lot to get my head around.

First, I must point out that the term "flag" is used to refer only to those parts of a Sagittal symbol that are not shafts or diacritics.

Noted, sorry for that.

{in }

That's not actually a diacritic. It's a flag -- the left scroll (left concave flag) or 19-schisma flag. It looks like a diacritic when it appears in conjunction with a left barb (left straight flag). However the true diacritics never actually touch the core symbol or shaft that they appear with.

Ah! That makes a lot of sense now, I don't think I ever saw the explanation for that symbol and was wondering a little about it. Great.


I know I might be in the minority, but I still stand by "pao-sharp" and "tao-flat", rather than the opposite ordering. My reasoning is this:
The whole thing IS the accidental. In pure Sagittal, it's one symbol. In mixed Sagittal, you're really modifying your nominal (G or D or whatever) by the whole accidental, and not necessarily "thinking G# and playing a comma lower".

In Ben Johnston's notation, 13:8 is notated as 8:5 raised by a "13" accidental on top. A 13:8 above C would be A13b, and I very much doubt you'd be calling it "A flat thirteen", but rather "A thirteen flat". 7:5 is notated as 36:25 lowered by a "7" accidental, and so 7:5 from C would be G7b, again, you wouldn't think people would be calling this "G flat seven"...

I realise these are both also chord names and there's that issue of not referring to a different entity here where there might not be in your notation. Still, I find that we should treat these accidentals as single entities. The Sagittal accidental for 24:25 down is a "phai-flat" (I think we agree on naming here). But why wouldn't sticking that on a G or B or whatever give you "G phai-flat" or "B phai-flat"?

to many musicians, "E flat" is just another note name in its own right, not really considered as a modification of an E. And so "E flat pai" is read as "(E flat) pai" and therefore it would be completely unnatural to insert the word "pai" between the "E" and the word "flat".

I see what you're saying, but this sounds like you're implying that E pai-flat (or E flat pai) is a modification of E flat, and not a note in its own right... I'd say many microtonalists would see them as "different notes" rather than just a "modification". I'd much rather think of pai-flat being a modification of FLAT, because what you're really changing is the accidentals. If that makes sense.
Also, I don't much like the idea of having a different rule for mixed and pure symbol name-order. So if it works in pure, I'd say we keep the same names for both.

As for the schisminas,

we can't just add a bare "o" or "i" to the end of the word

I'm not so sure. I could see glottal stops working, and the notation of such stops in most languages is with '. I'd say that's a VERY good notation for your little diacritics. The same way you can say and write "Hawai'i", you can say and write "tai'i" or "pao'o", or whatever.
Then again, some speakers might find this hard? I might have to do some research on glottal stop pronunciation (even when not a native phoneme) in a few languages. I know Japanese never natively uses them (and probably Maori too) but might be easy enough for speakers to pick up, easier than a real new phoneme.

Good to hear about wai and wao! It seemed too good an opportunity to pass up. And I'm very much liking some of the short names I'm seeing in George's post, though it'll take me a while to try them all out!

Probably get back to you in a bit!

[Dave Keenan]

Dave, Wow, there's quite a lot to get my head around.

That's always a problem when presenting anything that covers the whole Sagittal system. The initial impression is one of complexity. George and I call it "the freakout factor". In reality most people will never need to go beyond the Spartan set (the first 8). And the important thing to note is that it all comes down to recognising only 4 different shapes, in left-hand and right-hand varieties, attached to arrows pointing up or down. All you really have to learn is this:

p   k       5-comma                                    55-comma
j   t       7:11-comma (~13:17-S-diesis, ~29-S-diesis) 7-comma
r   n       19-schisma                                 5:7-kleisma (~11:13-kleisma)
s   z       17-kleisma                                 23-comma

And in fact most people will never have to deal with r, s or z.

The other consonants are optional. They merely allow one syllable names (if the language allows them distinct pronunciations), for common symbols that already have two-syllable names as combinations of the above.

f           25-S-diesis (~5:13-S-diesis)
g           35-M-diesis (~13-M-diesis, ~125-M-diesis)
v           11-M-diesis
w           11-L-diesis
d           35-L-diesis (~13-L-diesis, ~125-L-diesis)

[Dave Keenan]

This is a proposal I made in email to George about two months ago. It does not involve any changes to the Sagispeak pronunciations, only to 3 of the unpronounced ASCII shorthand characters:

I note that about 2/3 of the way through viewtopic.php?f=2&t=8 Cam Taylor apparently assumed that since and are nao and nai, "n" must be the downward ASCII shorthand, and "u" was the obvious choice for the corresponding upward ASCII shorthand. I'd be quite happy to accommodate that.

That means finding different ASCII shorthand characters for the rarely-used Athenians sanao and sanai , for which we currently have "u" down and "e" up. In that case we'd only have the characters "c e l o @" from which to choose. The characters "o" and "@" are too convex to be used for those. And we can't use a character for the down symbol that would clash with existing down symbols if people tried to pronounce it. That eliminates "c" and "l" as they would clash with "k" and "r", leaving "e" as the only possible down symbol. "c" would be the obvious partner based on their appearance, although neither "e" nor "c" bear any resemblance to the actual symbols and , but then neither do the current choices "u" and "e".

I include "w", "m" and "r" in the table below (despite there being no change to them), because George refers to them in his response. The proposed change is effectively a "clockwise rotation" of the characters "c", "e" and "u" in the table below.

We currently have:

Plain-text
shorthand      Sagispeak consonant(s)
---------------------------------------
Spartan symbols:
n  c      	n
w  m     	jat or w
Non-spartan single-flag symbols:
r  ;      	r
Remaining athenian symbols:
u  e     	san

The proposal is:

Plain-text
shorthand      Sagispeak consonant(s)
---------------------------------------
Spartan symbols:
n  u      	n
w  m     	jat or w
Non-spartan single-flag symbols:
r  ;      	r
Remaining athenian symbols:
e  c     	san

George's email response was:

While "n" and "u" are indeed opposites (that we used before as such), it's illogical and potentially confusing to make "n" down and "u" up when "m" is up and "w" is down. Also, "u" looks more like than , which is unfortunate when you consider that "r" already looks more like than . On the other hand, it's fairly easy to remember that the "c" character represents |(, [which represents ] which is a good reason not to change it.

As for "e" and "c" pair for , the "e" looks like it should be the up character, but we can't make "c" the down character, because it's a consonant. While the "c" does match the right flag, it's not as easy to remember that it would represent , as opposed to , as we presently have it.

So I can't agree with this proposal.

I suspect George meant to say: "u" looks more like than .

[Dave Keenan]

I note that it is extremely unlikely that anyone will ever use any ASCII shorthand for or because (a) they represent a 17-comma in JI and are used in only very large EDOs, and (b) in either proposal, "u e" or "e c", the mapping is too obscure, so people will almost certainly just use the ASCII longhand ~!( and ~|( instead.

And I note that in either proposal, we already do have "n" as downward while "m" is upward. We have already accepted that, just as we have accepted the similar problem George points out involving "r", in the interests of being consistent between pronunciation and downward shorthand. So having accepted that, I see no further problem in having "u" as upward while "w" is downward.

If in doubt, a user can reason as follows: "n" and "u" are obviously an up/down, or down/up, pair. The downward letter always corresponds to the consonant in the Sagispeak pronunciation (whenever the corresponding symbols have a single-consonant pronunciation, as these do). "u" is not a consonant, therefore "n" is down and "u" is up.

So we really should be asking only which ASCII shorthand is more logical, memorable or visually obvious:

n  c      

or

n  u      

[Xen-Gedankenwelt]

While I agree that it is confusing that n/u is used "in an upside-down way" compared to w/m, I think it becomes very intuitive if you imagine the Sagittal symbol / add the "stem" to the letter. If you think of the middle-line in 'm' or in a more round variant of 'w' (similar to ω) as the stem, the direction becomes intuitive. Similarly, you can think of 'n' as a reduced 'h', and of u as a reduced 'μ'.
(this may not be perfectly intuitive, since h already represents another accidental, but maybe it still helps)

I guess it's not perfect, but it'll work that way. I have to admit though that I didn't work much with Sagittal in practice yet, so my opinion may not carry much weight.