Template talk:pi-sc

Parameters tr and pos

What legitimate purposes do these currently undocumented parameters added by @SodhakSH in May 2021 serve? By their names, they should be for transliteration and part of speech, but transliteration is now provided for in the headword line. @Benwing2 pointed out that the transliteration in the headword is just clutter if this template it is used, so it is disabled by default for Pali-specific headword templates, but can be overridden. (This facility was not available in May 2021.) The part of speech is given by the header line. @Octahedron80 --RichardW57 (talk) 09:26, 3 July 2021 (UTC)

The transliteration is needed if it is different from the Roman equivalent, as occasionally happens even outside the Thai and Lao scripts, or the same spelling can have different transliterations depending on the writing system used. Two words with different transliterations should have different headword lines.

I have seen SodhakSH make dubious use of |pos= to introduce a subclassification as 'place name', but there are other mechanisms for that, and this trick is not sanely available for the main entries, which are in the Roman script. --RichardW57 (talk) 09:26, 3 July 2021 (UTC)

I propose that we create a temporary category to find these parameters' users, and then eliminate them. --RichardW57 (talk) 09:26, 3 July 2021 (UTC)

Repinging @Octahedron80 because of typo. --RichardW57 (talk) 09:30, 3 July 2021 (UTC)

Variant Options

@Octahedron80, SodhakSH, Fenakhay, 咽頭べさ, Mahagaja, AryamanA: Shouldn't variants be indicated by use of {{label}}?

The scheme also runs contrary to the changes SodakhSH and I have made to reduce the massive scope for human error when adding a word with homographs in other languages. At present one can just type {{pi-sc}} and usually get it right. (The Latin equivalent can often go wrong in Lao, infrequently in Thai, and rarely elsewhere.)

If we are going to do this, the labels are wrong or misleading. 'Traditional' and 'syllabic' are not fit for describing two classes of writing system, which are an abugida and an alphabet respectively. What is the evidence that writing as an abugida is the traditional system for the Lao script? The traditional way of writing Pali in Laos is to use the Lanna script. Writing Pali in an abugida using the Thai script is so 20th century; the traditional way of writing Pali in much of Thailand is to use the Khmer script, in the variant known as the Khom script.

For the Thai and Lao scripts, the default variant should be unspecified. There are quite a few words where a common inflection table is used for the two different systems. No-one had objected to this until I discovered and pointed out that it did not work with transliteration for masculine and neuter nouns. (The generation of the inflections could be restructured to keep track of the class of writing system.) Octahedron80 is facing the task of classifying all the Thai and Lao script lemmas, and he needs maintenance categories to know which still need to be classified. By default, most Lao script entries are wrong if 'Laoo' is going to identify the abugidic system - most Lao script entries are alphabetic, not abugidic.

If one's going to keep track of the writing system, there are other divisions that would also be appropriate for the Lao script. Does the writing system use the Buddhist Institute additions/restorations? Is the nukta-based system to be treated as a third way? What of subtler variations, such as the handling of occlusive codas and of consonant clusters starting with ສ or ຕ? (The initial consonant of the cluster may be crowned with a cancellation mark or possibly an unencoded Lao clone of yammakkan.) There are also differences in repertoire - are ຣ and ຢ used? (I already support declension of Lao nouns with stems in -ັລ, corresponding to nouns in -ar.)

The distinction between the Burmese and Mon variants for the Burmese script seem less prominent than the differences between Northern Thai and Lao Lanna script Pali. --RichardW57 (talk) 07:14, 5 July 2021 (UTC)

I'm adding categories cat:pi-sc-generic and cat:pi-sc-specific to help complete or reverse Octahedron80's change pending what the consensus is. I'm against it. @Octahedron80, SodhakSH, Fenakhay, 咽頭べさ, Mahagaja, AryamanA --RichardW57m (talk) 11:48, 5 July 2021 (UTC)

@Octahedron80: To cover automatic script detection, I'm going to have to restructure to identify the script etc. and then invoke a separate template to format the identification of the script etc. The other ways seem too difficult to maintain. --RichardW57 (talk) 19:37, 5 July 2021 (UTC)  Done

At Thai Wiktionary, the translation of these texts just makes sanse. I didn't think of better word to describe between two systems. The system 2 is syllabic; it does describe itself. (If someone wants to insert dashes inbetween a word, the word will become complete syllables.) The system 1 is traditional(?) for its written script, not for other script that already has its entry (Khmr & Lana). When system 1 & 2 results the same, just double the # list; it should not be hidden. This is good for wandering reader to know what type(s) the word is.

Both alphabetic and abugidic writing systems are (generally) syllabic - just remember that the short Pali letter names are not the same as the Thai and Lao names. Doubling the number of senses might work for the Thai script because the senses are cross-references to the Latin script equivalent. However, that could get nasty if there is a local meaning - I am not sure that such a meaning will necessarily belong to the Latin script form. In general, doubling without a difference looks wrong. For identical forms derived from words with different spellings in the various writing systems, I'd started to merge the senses because that seemed better. It doesn't work well with our current set of templates. --RichardW57m (talk) 11:06, 6 July 2021 (UTC)

For the Lao script, your argument will potentially lead to roughly half-a-dozen lines. I need to review the writing systems, but for Lao I think we have:

1. Buddhist Institute abugida. Should it be split for the option of yamakkan in the -sm- inflections and absolutives?
2. Buddhist Institute alphabet. Potentially also a split for the option of yamakkan.
3. Nukta-based alphabet (rare - not Windows XP compatible)
4. Lao repertoire, using yo yaa, but not otherwise modifying consonants beyond the mergers.
5. Lao repertoire, using yo yung, but not otherwise modifying consonants beyond the mergers.
6. Lao repertoire, using Lao spelling rules.

The last three may need to split according to whether they use yamakkan, the cancellation mark, or nothing, for clusters at the start of Pali syllables, with its not very 'syllabic' behaviour in the middle of words. An exemplifying word for these systems is ສ໌ວາກຂາຕະ (svākkhāta).

Rather than one sense per non-Roman writing system, it would be better to tag entries with a list of applicable writing systems. --RichardW57m (talk) 11:06, 6 July 2021 (UTC)

About Lao script, the proposal to encode new characters said

"To enable the Lao script to completely and faithfully express the Pali language and also write Indic loan words with proper etymological spelling, Buddhist scholars tried to implement a script reform. In the 1930s Buddhist scholar Maha Sila Viravong at the initiation of the Buddhist Institute in Vientiane and approved by the Buddhist Academic Council, added an additional set of characters to support Pali (and also Sanskrit) by filling in the missing gaps. The Buddhist Institute using the expanded script published several books such as Dhammapada and Pali Grammar. Their aim was to make Pali more accessible to the common public. But the addition met with little widespread support and finally by 1975, these additional characters were mostly out of use. There is a revived interest in the characters and a few modern publications have been printed using them in an attempt to propose an etymological orthography for Lao."

So the "traditional" could be replaced with the "Buddhist Institute" or something similar. About Thai script, I have no idea which term to replace traditional. Both systems are generally used in publications. They can be seen anywhere, in textbooks (ex1), official documents, or at the shrines. --Octahedron80 (talk) 00:26, 6 July 2021 (UTC)

Well, snappy words are 'abugidic' and 'alphabetic'. The only drawback is that a lot of people seem to confuse the concept of abugida with that of alphasyllabary. For inflection tables, I've taken to labelling them as 'with implicit vowels' and 'without implicit vowels' when I've felt the need to separate them. --RichardW57m (talk) 11:06, 6 July 2021 (UTC)

@Octahedron80: I've changed the wording to use 'with implicit vowels' and 'without implicit vowels' for the Thai script in the cases where the two forms are distinguished. --RichardW57m (talk) 13:04, 26 August 2021 (UTC)

And also don't forget that Myanmar script has 4 variations. Burmese and Mon Pali may be very similar but there are some distinguishable points, whereas old and new Shan Pali are completely different. (I have seen Shan Pali textbooks.) --Octahedron80 (talk) 00:38, 6 July 2021 (UTC)

More than that. Don't forget Thai Mon Pali. It uses a stack rather than Great Sa for the genitive singular masculine etcetera. There's also Khamti, which may come with its own set of variations. I'm strongly tempted to treat Shan as a separate script. However, that may look bad when a word is the same in both writing systems - we might need a description such as 'Burmese and Shan script form of'. Remember that Module:pi-Latn-translit doesn't need to worry about forms in different writing systems being the same, and that it needs significant manual supplementation for the Lanna and Lao scripts. We can't generate an inflection table just by transliterating a Roman script inflection table. (You may remember that there was a time when generating inflection tables corrupted the stem, e.g. for ᨻᩮᩣᨵᩥ (bodhi).) On the other hand, when different writing systems have different sets of inflections, there is a case for making the word into multiple (subordinate) lemmas. That is what I do when the senses have different genders, e.g dhātu. --RichardW57m (talk) 11:06, 6 July 2021 (UTC)

"Old Shan" isn't very old. "Stacking Shan" and "Non-Stacking Shan" might be better names. --RichardW57m (talk) 11:06, 6 July 2021 (UTC)

I didn't see any Thai-Mon Pali text yet. I only saw some weird letters in signs from Intubesa showing me. (Yes, he ever visited Thai-Mon temples) He said Thai-Mon Pali is currently not able to write because the weird letters are not encoded, that look like Old Mon script. --Octahedron80 (talk) 14:18, 6 July 2021 (UTC)

There's a report on it at L2/20-163; there's some easily readable Thai Mon Pali on p9 and some harder to read stuff on p12. One can basically write it now, though one may need a font that doesn't necessitate the use of TALL AA. That's a similar issue to Tai Tham - some Tai Tham fonts need one to write TALL AA for text to be readable, which causes problems for writing systems that don't use it. I don't know if there is any prospect of 'WIGGLY AA' being encoded for the non-ligating Thai Mon round aa. It looks as though there might be some Pali in a Tai Laing writing systems within the Burmese script. That has a convention that might break transliteration, but nothing major. I found some Tai Tham that breaks transliteration at ᨾᩉᩣᨻ᩠ᨵᩮᩥᩣ, but I'll fix that tonight - I just need to apply |subst= to the inflection table. --RichardW57m (talk) 15:15, 6 July 2021 (UTC)

Alerting @Apisite. --RichardW57m (talk) 12:20, 6 July 2021 (UTC)

|id=

@Apisite, Octahedron80 I intend to add this parameter, for which see {{senseid}} and {{etymid}}, some time this week. I don't intend to do more than a few test examples this week. I think non-Roman forms of santa would be good examples to use. The coding change will be a very small change to the current coding of the template. --RichardW57m (talk) 16:53, 3 July 2023 (UTC)

I don't think thwikt need that however. --Octahedron80 (talk) 08:37, 4 July 2023 (UTC)

Does thwikt use {{senseid}}? RichardW57 (talk) 21:04, 4 July 2023 (UTC)

I've now implemented and tested them. Now all we need is to stop collapsing tables messing everything up. --RichardW57 (talk) 21:03, 4 July 2023 (UTC)