poetic license grammar

[kylebgorman]

Have you found instances of synizesis and/or diaeresis in the data? I.e., are there lines that only scan once these rules are added in?

[jillianchang]

I found one example of synizesis (I put it as the last test). Haven’t found any diaeresis.

[kylebgorman]

I found one example of synizesis (I put it as the last test). Haven’t found any diaeresis.

Awesome.

[kylebgorman]

Not off the top of my head, no, sorry.

K

On Tue, Jul 27, 2021 at 8:28 PM jillianchang @.***> wrote:

@.**** commented on this pull request.

In grammars/optional.grm https://github.com/CUNY-CL/LatinScansion/pull/18#discussion_r677887771:

@@ -0,0 +1,45 @@ +# Implements sandhi phenomena, such as elision, resyllabification, +# synezisis, hypermetric lengthening, etc.

Can you think of any examples of diaeresis involving the [j] glide?

— You are receiving this because you commented.

Reply to this email directly, view it on GitHub https://github.com/CUNY-CL/LatinScansion/pull/18#discussion_r677887771, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABG4OKOEPEBONU3AP4XSG3TZ5FJLANCNFSM5BBJ767Q .

[jillianchang]

Found something sorta strange. When testing the line "iːtaliãː faːtoː profugus laːwiːniakwe weːnit" in the optional grammar file for synizesis, the test passes. But when testing SCAN using the pipes, terminal throws "Composition Failure." This is weird since I know that synizesis should allow the line to be scanned (I tested the line with the synizesis modified pronunciation against meter.grm)

Huh, that's not obvious to me why. Maybe you can give me a command to replicate? The synizesis is in laːwiːnjakwe, right?

[jillianchang]

Can a line have more than one optional rule applied to it?

[kylebgorman]

Can a line have more than one optional rule applied to it?

Yes, that should work fine. If R1 and R2 are both optional rules, R1 @ R2 should allow both rules to apply zero, one, two, … etc times. And both R1 and R2 can apply to the same line, if the conditions in the rule are met and if apply R1 doesn’t prevent the conditions of R2 from being met.

—

You are receiving this because you commented.

Reply to this email directly, view it on GitHub https://github.com/CUNY-CL/LatinScansion/pull/18#issuecomment-888618393, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABG4OKWNLT2QVMY6J2QOOTT2BV3FANCNFSM5BBJ767Q .

[jillianchang]

The meter only works for the line "impulerit tantajne animiːs kajlestibus iːraj" (which is "impulerit tantajnanimiːs kajlestibu siːraj" with resyllabification applied) if "kajlestibu" is [kaj.les.ti.bu] instead of [kaj.le.sti.bu]. In other words, the consonant cluster is split instead of being the onset. Is this some sort of optional rule?

[kylebgorman]

The meter only works for the line "impulerit tantajne animiːs kajlestibus iːraj" (which is "impulerit tantajnanimiːs kajlestibu siːraj" with resyllabification applied) if "kajlestibu" is [kaj.les.ti.bu] instead of [kaj.le.sti.bu]. In other words, the consonant cluster is split instead of being the onset. Is this some sort of optional rule?

I agree, that second syllable in caelestibus has to be heavy---the head of a foot---or it won't work.

I just checked Pharr to make sure I didn't make a typo here, and no, this is correct.

This is a really cool observation. I suspect that word-internal sC clusters (where C is a consonant) do in general come out as [s.C], and we should instead fix this in meter.grm's definition of SYLLABLE then, in a separate PR. (I think then the generalization might have to be slightly different for initial and non-initial syllables---you would want to allow a word initial sC to be an onset but split a medial sC into coda and onset.

(In English this is variable: some people prefer A.las.ka and others A.la.ska. You can test this for yourself by doing the following: say every syllable of the word but say it twice. I say uh-uh-las-las-ka-ka, but some people say uh-uh-la-la-ska-ska.)

[jillianchang]

Found something sorta strange. When testing the line "iːtaliãː faːtoː profugus laːwiːniakwe weːnit" in the optional grammar file for synizesis, the test passes. But when testing SCAN using the pipes, terminal throws "Composition Failure." This is weird since I know that synizesis should allow the line to be scanned (I tested the line with the synizesis modified pronunciation against meter.grm)

Huh, that's not obvious to me why. Maybe you can give me a command to replicate? The synizesis is in laːwiːnjakwe, right?

Correct, that's where the synizesis is. What command do you mean?

[kylebgorman]

I wanted the command you’re running to get the error; I was going to replicate locally.

K

On Fri, Jul 30, 2021 at 10:50 AM jillianchang @.***> wrote:

Found something sorta strange. When testing the line "iːtaliãː faːtoː profugus laːwiːniakwe weːnit" in the optional grammar file for synizesis, the test passes. But when testing SCAN using the pipes, terminal throws "Composition Failure." This is weird since I know that synizesis should allow the line to be scanned (I tested the line with the synizesis modified pronunciation against meter.grm)

Huh, that's not obvious to me why. Maybe you can give me a command to replicate? The synizesis is in laːwiːnjakwe, right?

Correct, that's where the synizesis is. What command do you mean?

— You are receiving this because you commented.

Reply to this email directly, view it on GitHub https://github.com/CUNY-CL/LatinScansion/pull/18#issuecomment-889944831, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABG4OK3KKOQOWFW3W2ZQBDT2K34FANCNFSM5BBJ767Q .

[jillianchang]

Okay, my bad.

cat /Users/jillianchang/LatinScansion/data/Aeneid/Aeneid01.txt | ./rewriter.py --far /Users/jillianchang/LatinScansion/grammars/normalize.far --rules NORMALIZE | ./rewriter.py --far /Users/jillianchang/LatinScansion/grammars/pron.far --rules PRON | ./rewriter.py --far /Users/jillianchang/LatinScansion/grammars/optional.far --rules OPTIONAL | ./rewriter.py --far /Users/jillianchang/LatinScansion/grammars/meter.far --rules SCAN

On Jul 30, 2021, at 10:52 AM, Kyle Gorman @.***> wrote:

I wanted the command you’re running to get the error; I was going to replicate locally.

K

On Fri, Jul 30, 2021 at 10:50 AM jillianchang @.***> wrote:

Found something sorta strange. When testing the line "iːtaliãː faːtoː profugus laːwiːniakwe weːnit" in the optional grammar file for synizesis, the test passes. But when testing SCAN using the pipes, terminal throws "Composition Failure." This is weird since I know that synizesis should allow the line to be scanned (I tested the line with the synizesis modified pronunciation against meter.grm)

Huh, that's not obvious to me why. Maybe you can give me a command to replicate? The synizesis is in laːwiːnjakwe, right?

Correct, that's where the synizesis is. What command do you mean?

— You are receiving this because you commented.

Reply to this email directly, view it on GitHub https://github.com/CUNY-CL/LatinScansion/pull/18#issuecomment-889944831, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABG4OK3KKOQOWFW3W2ZQBDT2K34FANCNFSM5BBJ767Q .

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/CUNY-CL/LatinScansion/pull/18#issuecomment-889946360, or unsubscribe https://github.com/notifications/unsubscribe-auth/ANYD7FZYJNXBWZIAZKPXXADT2K4EVANCNFSM5BBJ767Q.

[kylebgorman]

Not really, since they’re mandatory.

K

On Fri, Jul 30, 2021 at 11:09 AM jillianchang @.***> wrote:

@.**** commented on this pull request.

In grammars/optional.grm https://github.com/CUNY-CL/LatinScansion/pull/18#discussion_r679999983:

export RESYLLABIFY = Optimize[

 CDRewrite[u.Insert[" "], "", i.CONSONANT " " i.PHONEMIC_VOWEL, sigma_star, 'ltr', 'opt'] @

 CDRewrite[u.Delete[" "], " " i.CONSONANT, i.PHONEMIC_VOWEL, sigma_star]

];

-# TODO: orthographically annotate using "‿".

+

+# The consonantization of i and u into their corresponding glides, [j] and [w].

+export SYNIZESIS = Optimize [

• CDRewrite[("i" : "j") |

• ("u": "w"), "", i.SHORT_VOWEL, sigma_star, 'ltr', 'opt']

+];

+

Are these rules (the "eu", "ou", "ui", "ei" diphthongs that are normally pronounced as hiatuses) considered synizesis?

— You are receiving this because you commented.

Reply to this email directly, view it on GitHub https://github.com/CUNY-CL/LatinScansion/pull/18#discussion_r679999983, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABG4OK6XHJ3E4DK6CWN7ITT2K6D5ANCNFSM5BBJ767Q .

[jillianchang]

I meant if we discover other words that use those rules to make the meter.

On Jul 30, 2021, at 11:11 AM, Kyle Gorman @.***> wrote:

Not really, since they’re mandatory.

K

On Fri, Jul 30, 2021 at 11:09 AM jillianchang @.***> wrote:

@.**** commented on this pull request.

In grammars/optional.grm https://github.com/CUNY-CL/LatinScansion/pull/18#discussion_r679999983:

export RESYLLABIFY = Optimize[

CDRewrite[u.Insert[" "], "", i.CONSONANT " " i.PHONEMIC_VOWEL, sigma_star, 'ltr', 'opt'] @

CDRewrite[u.Delete[" "], " " i.CONSONANT, i.PHONEMIC_VOWEL, sigma_star]

];

-# TODO: orthographically annotate using "‿".

+

+# The consonantization of i and u into their corresponding glides, [j] and [w].

+export SYNIZESIS = Optimize [

• CDRewrite[("i" : "j") |

• ("u": "w"), "", i.SHORT_VOWEL, sigma_star, 'ltr', 'opt']

+];

+

Are these rules (the "eu", "ou", "ui", "ei" diphthongs that are normally pronounced as hiatuses) considered synizesis?

— You are receiving this because you commented.

Reply to this email directly, view it on GitHub https://github.com/CUNY-CL/LatinScansion/pull/18#discussion_r679999983, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABG4OK6XHJ3E4DK6CWN7ITT2K6D5ANCNFSM5BBJ767Q .

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/CUNY-CL/LatinScansion/pull/18#issuecomment-889958091, or unsubscribe https://github.com/notifications/unsubscribe-auth/ANYD7F3DZITDEJYACBDEWGTT2K6ILANCNFSM5BBJ767Q.

[kylebgorman]

Oh I understand now. If you find something like that, I guess I would call it synizesis too.

K

On Fri, Jul 30, 2021 at 11:18 AM jillianchang @.***> wrote:

I meant if we discover other words that use those rules to make the meter.

On Jul 30, 2021, at 11:11 AM, Kyle Gorman @.***> wrote:

Not really, since they’re mandatory.

K

On Fri, Jul 30, 2021 at 11:09 AM jillianchang @.***> wrote:

@.**** commented on this pull request.

In grammars/optional.grm < https://github.com/CUNY-CL/LatinScansion/pull/18#discussion_r679999983>:

export RESYLLABIFY = Optimize[

CDRewrite[u.Insert[" "], "", i.CONSONANT " " i.PHONEMIC_VOWEL, sigma_star, 'ltr', 'opt'] @

CDRewrite[u.Delete[" "], " " i.CONSONANT, i.PHONEMIC_VOWEL, sigma_star]

];

-# TODO: orthographically annotate using "‿".

+

+# The consonantization of i and u into their corresponding glides, [j] and [w].

+export SYNIZESIS = Optimize [

• CDRewrite[("i" : "j") |

• ("u": "w"), "", i.SHORT_VOWEL, sigma_star, 'ltr', 'opt']

+];

+

Are these rules (the "eu", "ou", "ui", "ei" diphthongs that are normally pronounced as hiatuses) considered synizesis?

— You are receiving this because you commented.

Reply to this email directly, view it on GitHub < https://github.com/CUNY-CL/LatinScansion/pull/18#discussion_r679999983>, or unsubscribe < https://github.com/notifications/unsubscribe-auth/AABG4OK6XHJ3E4DK6CWN7ITT2K6D5ANCNFSM5BBJ767Q

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/CUNY-CL/LatinScansion/pull/18#issuecomment-889958091>, or unsubscribe < https://github.com/notifications/unsubscribe-auth/ANYD7F3DZITDEJYACBDEWGTT2K6ILANCNFSM5BBJ767Q .

— You are receiving this because you commented.

Reply to this email directly, view it on GitHub https://github.com/CUNY-CL/LatinScansion/pull/18#issuecomment-889962965, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABG4OKHCP7ZXEYHV7NEGPTT2K7C7ANCNFSM5BBJ767Q .

[kylebgorman]

There's something wrong with this Makefile. That thraxmakedep tool doesn't understand how to deal with the case when a single directory has multiple "top-level" grammars (ones we want FARs for), like we have here. I can hack this later, or if you're interested, you can research Makefile syntax (but it's a bit obtuse).

[kylebgorman]

Okay I see why applying the SYNIZESIS rule didn't work with SCAN. The reason is that the grammar picks (arbitrarily) one of the two possibilities: applying synizesis, or not. (Why this is arbitrary is discussed in the rewrite chapter of FSTP.) If you add the --one-top-rewrite flag to the synizesis stage of rewriting, you'll see why: it raises an error indicating that there are multiple ways to rewrite this (and other) input.

What we need then is to combine the optional rules (somehow) with the meter rules, so they're applied only when they help scan the line. And this is not something we can do using rewriting tool (since it only prints one output line per input line): we have to sit down and figure out how to combine the grammars to express that idea, which we can with weighted FSTs. We can take a stab at that Monday.

[jillianchang]

Sure, we can talk about it on Monday. Is --one-top-rewrite a flag for the Python rewrite tool?

[kylebgorman]

Yes that’s right.

On Sun, Aug 1, 2021 at 5:06 PM jillianchang @.***> wrote:

Sure, we can talk about it on Monday. Is --one-top-rewrite a flag for the Python rewrite tool?

— You are receiving this because you commented.

Reply to this email directly, view it on GitHub https://github.com/CUNY-CL/LatinScansion/pull/18#issuecomment-890586607, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABG4OKOP5KP7R67EIDALUDT2WZNFANCNFSM5BBJ767Q .

[jillianchang]

How can I debug FINAL_SCAN tests that fail?

[kylebgorman]

Other than using the intermediate rules I can’t think of anything straightforward. Maybe I can build something…

On Wed, Aug 4, 2021 at 4:57 PM jillianchang @.***> wrote:

How can I debug FINAL_SCAN tests that fail?

— You are receiving this because you commented.

Reply to this email directly, view it on GitHub https://github.com/CUNY-CL/LatinScansion/pull/18#issuecomment-892968596, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABG4OJ2P36WNFEIV2WZTJTT3GSR3ANCNFSM5BBJ767Q . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email .

[jillianchang]

Note: I created a new file called optional_tests.grm which is basically a copy of optional.grm (but without the weights) and I use it to run individual tests. I then removed those tests from optional.grm, since the tests don't work when weights are added. This is probably not the greatest practice since I sort of have two files of the same thing (and so if I edit one file, I'll end up having to make the same change to the other), so let me know if you have a better idea.

[kylebgorman]

This looks like it's coming together---let me know what else you need.

[jillianchang]

Do you have to create a base branch on CUNY-CL called optional for me to make a PR to that?

On Aug 5, 2021, at 10:35 AM, Kyle Gorman @.***> wrote:

@kylebgorman commented on this pull request.

In grammars/final_scan.grm https://github.com/CUNY-CL/LatinScansion/pull/18#discussion_r683513820:

];

Tests h deletion + resyllabification, and another resyllabification.

-final_scan_8 = AssertEqual[

• "hiːk kurrus fuit hok reːŋnũː dea gentibus esse" @ OPTIONAL,
• "hiːk kurrus fui tok reːŋnũː dea gentibu sesse" +final_scan_9 = AssertEqual[
• "hiːk kurrus fuit hok reːŋnũː dea gentibus esse" @ FINAL_SCAN,
• "SDSDDT" +];

• +## Tests final pronounciation outputs. +final_pron_1 = u.AssertInputEqual[

• "wiː superũː sajwaj memorẽː juːnoːnis ob iːrãː",
• FINAL_SCAN,
• "wiː superũː sajwaj memorẽː juːnoːni so biːrãː" ]; Oh dang, I see why.

Let me suggest something: take these tests and submit them as a PR to a non-master branch. (I.e., you'll issue a PR from jillianchang:optional to CUNY-CL:optional.) Then I can work them into a Python unit testing frameworking. Figuring out how to do this properly is always hard as heck for me and may take me a few hours...

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/CUNY-CL/LatinScansion/pull/18#discussion_r683513820, or unsubscribe https://github.com/notifications/unsubscribe-auth/ANYD7F5OJZJ5QH5LVWAHNEDT3KOUBANCNFSM5BBJ767Q. Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email.

[kylebgorman]

I don't think so but let me know if you get stuck.

On Thu, Aug 5, 2021 at 12:07 PM jillianchang @.***> wrote:

Do you have to create a base branch on CUNY-CL called optional for me to make a PR to that?

On Aug 5, 2021, at 10:35 AM, Kyle Gorman @.***> wrote:

@kylebgorman commented on this pull request.

In grammars/final_scan.grm < https://github.com/CUNY-CL/LatinScansion/pull/18#discussion_r683513820>:

];

Tests h deletion + resyllabification, and another resyllabification.

-final_scan_8 = AssertEqual[

• "hiːk kurrus fuit hok reːŋnũː dea gentibus esse" @ OPTIONAL,
• "hiːk kurrus fui tok reːŋnũː dea gentibu sesse" +final_scan_9 = AssertEqual[
• "hiːk kurrus fuit hok reːŋnũː dea gentibus esse" @ FINAL_SCAN,
• "SDSDDT" +];

• +## Tests final pronounciation outputs. +final_pron_1 = u.AssertInputEqual[

• "wiː superũː sajwaj memorẽː juːnoːnis ob iːrãː",
• FINAL_SCAN,
• "wiː superũː sajwaj memorẽː juːnoːni so biːrãː" ]; Oh dang, I see why.

Let me suggest something: take these tests and submit them as a PR to a non-master branch. (I.e., you'll issue a PR from jillianchang:optional to CUNY-CL:optional.) Then I can work them into a Python unit testing frameworking. Figuring out how to do this properly is always hard as heck for me and may take me a few hours...

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/CUNY-CL/LatinScansion/pull/18#discussion_r683513820>, or unsubscribe < https://github.com/notifications/unsubscribe-auth/ANYD7F5OJZJ5QH5LVWAHNEDT3KOUBANCNFSM5BBJ767Q . Triage notifications on the go with GitHub Mobile for iOS < https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android < https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email .

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/CUNY-CL/LatinScansion/pull/18#issuecomment-893580115, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABG4OIAEDUTZLGQYEQIHODT3KZMDANCNFSM5BBJ767Q . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email .

[jillianchang]

Is this how I do it?

On Aug 5, 2021, at 12:29 PM, Kyle Gorman @.***> wrote:

I don't think so but let me know if you get stuck.

On Thu, Aug 5, 2021 at 12:07 PM jillianchang @.***> wrote:

Do you have to create a base branch on CUNY-CL called optional for me to make a PR to that?

On Aug 5, 2021, at 10:35 AM, Kyle Gorman @.***> wrote:

@kylebgorman commented on this pull request.

In grammars/final_scan.grm < https://github.com/CUNY-CL/LatinScansion/pull/18#discussion_r683513820>:

];

Tests h deletion + resyllabification, and another resyllabification.

-final_scan_8 = AssertEqual[

• "hiːk kurrus fuit hok reːŋnũː dea gentibus esse" @ OPTIONAL,
• "hiːk kurrus fui tok reːŋnũː dea gentibu sesse" +final_scan_9 = AssertEqual[
• "hiːk kurrus fuit hok reːŋnũː dea gentibus esse" @ FINAL_SCAN,
• "SDSDDT" +];

• +## Tests final pronounciation outputs. +final_pron_1 = u.AssertInputEqual[

• "wiː superũː sajwaj memorẽː juːnoːnis ob iːrãː",
• FINAL_SCAN,
• "wiː superũː sajwaj memorẽː juːnoːni so biːrãː" ]; Oh dang, I see why.

Let me suggest something: take these tests and submit them as a PR to a non-master branch. (I.e., you'll issue a PR from jillianchang:optional to CUNY-CL:optional.) Then I can work them into a Python unit testing frameworking. Figuring out how to do this properly is always hard as heck for me and may take me a few hours...

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/CUNY-CL/LatinScansion/pull/18#discussion_r683513820>, or unsubscribe < https://github.com/notifications/unsubscribe-auth/ANYD7F5OJZJ5QH5LVWAHNEDT3KOUBANCNFSM5BBJ767Q . Triage notifications on the go with GitHub Mobile for iOS < https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android < https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email .

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/CUNY-CL/LatinScansion/pull/18#issuecomment-893580115, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABG4OIAEDUTZLGQYEQIHODT3KZMDANCNFSM5BBJ767Q . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email .

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/CUNY-CL/LatinScansion/pull/18#issuecomment-893596494, or unsubscribe https://github.com/notifications/unsubscribe-auth/ANYD7F3PFQBIQX5GC2CECETT3K36XANCNFSM5BBJ767Q. Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email.

[kylebgorman]

I think once you create the PR you can't change what branch it targets. I would suggest just moving the test stuff out of the PR for now. The optional grammar itself looks great.

[jillianchang]

Sorry, what commands should I do to do that?

On Aug 5, 2021, at 1:04 PM, Kyle Gorman @.***> wrote:

I think once you create the PR you can't change what branch it targets. I would suggest just moving the test stuff out of the PR for now. The optional grammar itself looks great.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/CUNY-CL/LatinScansion/pull/18#issuecomment-893618901, or unsubscribe https://github.com/notifications/unsubscribe-auth/ANYD7FYQCXN5BGYHWPJQYYLT3LAAXANCNFSM5BBJ767Q. Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email.

[kylebgorman]

This isn't the "right" way to do it but I think I'd make a (file) copy of the files you want to revert, store them somewhere, then remove them from the PR using git checkout.

Then, make a separate PR and copy the files you reverted back into it.

K

[kylebgorman]

Maybe we should add e-gliding with 10x the cost of I-gliding?

On Mon, Aug 9, 2021 at 4:50 PM jillianchang @.***> wrote:

@.**** commented on this pull request.

In grammars/optional.grm https://github.com/CUNY-CL/LatinScansion/pull/18#discussion_r685511267:

export RESYLLABIFY = Optimize[

 CDRewrite[u.Insert[" "], "", i.CONSONANT " " i.PHONEMIC_VOWEL, sigma_star, 'ltr', 'opt'] @

 CDRewrite[u.Delete[" "], " " i.CONSONANT, i.PHONEMIC_VOWEL, sigma_star]

];

-# TODO: orthographically annotate using "‿".

+

+# The consonantization of i and u into their corresponding glides, [j] and [w].

+export SYNIZESIS = Optimize [

• CDRewrite[("i" : "j") |

• ("u": "w"), "", i.SHORT_VOWEL, sigma_star, 'ltr', 'opt']

+];

+

Hmm, if it applied to that syllable I think you'd have to get [rja] or something like that. Definitionally you can't have two nuclei in a syllable though. Maybe that does in fact happen! The combined grammars will hopefully help us discover.

I think it actually does happen! aureā composuit spondā mediamque locāvit (1.698)

— You are receiving this because you modified the open/close state.

Reply to this email directly, view it on GitHub https://github.com/CUNY-CL/LatinScansion/pull/18#discussion_r685511267, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABG4OPG2BXKPRCDEBJQAMLT4A5PZANCNFSM5BBJ767Q . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email .

[jillianchang]

Yes, that sounds good. There is the problem that although we want it to be [au.rja], I think the meter grammar would make it [aur.ja] since rj isn't a valid consonant cluster. But I guess that doesn't matter for scansion purposes..?

[kylebgorman]

Yes, that sounds good. There is the problem that although we want it to be [au.rja], I think the meter grammar would make it [aur.ja] since rj isn't a valid consonant cluster. But I guess that doesn't matter for scansion purposes..?

It doesn't matter in this case because the preceding syllable is "long by nature" anyways. But it would have mattered if the preceding vowel was a short monophthong. Even though the way we use the variable grammar is filtered by the metrical grammars, we have not yet decided how to segment words into syllables wben we're applying the variable grammar, so all we need to do is allow [Cj onsets, I think.

[jillianchang]

Yes, that sounds good. There is the problem that although we want it to be [au.rja], I think the meter grammar would make it [aur.ja] since rj isn't a valid consonant cluster. But I guess that doesn't matter for scansion purposes..?

It doesn't matter in this case because the preceding syllable is "long by nature" anyways. But it would have mattered if the preceding vowel was a short monophthong. Even though the way we use the variable grammar is filtered by the metrical grammars, we have not yet decided how to segment words into syllables wben we're applying the variable grammar, so all we need to do is allow [Cj onsets, I think.

Wouldn't that pose a problem for words that have Cj anyway without the glide? (e.g. disjice is [dis.ji.ke], not [di.sji.ke])

[kylebgorman]

Good point. Let me think about it and get back to you.

On Mon, Aug 9, 2021 at 9:23 PM jillianchang @.***> wrote:

Yes, that sounds good. There is the problem that although we want it to be [au.rja], I think the meter grammar would make it [aur.ja] since rj isn't a valid consonant cluster. But I guess that doesn't matter for scansion purposes..?

It doesn't matter in this case because the preceding syllable is "long by nature" anyways. But it would have mattered if the preceding vowel was a short monophthong. Even though the way we use the variable grammar is filtered by the metrical grammars, we have not yet decided how to segment words into syllables wben we're applying the variable grammar, so all we need to do is allow [Cj onsets, I think.

Wouldn't that pose a problem for words that have Cj anyway without the glide? (e.g. disjice is [dis.ji.ke], not [di.sji.ke])

— You are receiving this because you modified the open/close state.

Reply to this email directly, view it on GitHub https://github.com/CUNY-CL/LatinScansion/pull/18#issuecomment-895658167, or unsubscribe https://github.com/notifications/unsubscribe-auth/AABG4OO5WSJTSG6WXLVAFK3T4B5Q3ANCNFSM5BBJ767Q . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&utm_campaign=notification-email .