Help talk:IPA/Lombard

Dialect-specific approximations

The English approximations are supposed to be helpful to the lay reader. Finding relatively obscure dialects that happen to feature the most phonetically accurate phone is not helpful. We should be finding the closest approximations in more general and well-known dialects. I began editing with this in mind but was reverted. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 17:28, 24 July 2018 (UTC)

I agree with that, and it's general practice to do so, but when sounds are so many I think it's better to leave them the most neatly distinct from one another so that readers understand useful differences in pronunciation. イヴァンスクルージ九十八_（会話） 19:02, 24 July 2018 (UTC)

@IvanScrooge98: Yeah, but that's the problem - when you choose obscure (or even not so obscure) dialects to approximate Lombard sounds (or sounds of any other language/dialect), it's of no help to laymen. They'd either have to have incredibly sensitive ears or speak that dialect themselves.

The column is titled English approximation, and there's a reason for that. You can only do so much when you compare the sounds of English to what you can hear in other languages. Kbb2 (ex. Mr KEBAB) (talk) 19:11, 24 July 2018 (UTC)

What I meant is, while "Welsh English" may not be straightforward, it would still remind "some specific British English" (considering this is somewhat more familiar), which is in my opinion much better than just removing the accent tags. イヴァンスクルージ九十八_（会話） 19:26, 24 July 2018 (UTC)

@IvanScrooge98: Actually /ɜː/ is rounded only in Southern Wales (the rest of Wales uses a rather RP-like realization) and Scottish /a/ varies from [æ] to [ɑ] depending on the region and speaker. So even these aren't specific enough. Also, native speakers of English can't really tell the difference between rounded and fronted [œː~øː] and raised [ɘː~ɨː] variants of NURSE. They sound about the same to them and they don't perceive them as very different to the RP-like [əː]. Kbb2 (ex. Mr KEBAB) (talk) 19:33, 24 July 2018 (UTC)

It is because they don't recur everywhere in Wales or Scotland I thought they were neither too specific to be misunderstood nor too general to mistake phonemes. イヴァンスクルージ九十八_（会話） 07:24, 25 July 2018 (UTC)

@IvanScrooge98: They definitely are too specific - see above. Kbb2 (ex. Mr KEBAB) (talk) 11:32, 25 July 2018 (UTC)

It's a fool's errand to try to use English approximations to convey phonemic contrasts that don't exist in English. Is that the reason for the revert? — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 20:54, 24 July 2018 (UTC)

Well, I am a fool (≧▽≦) However, as I said before I tried to give an idea of sound changes that do not occur in English (I mean, I wouldn't have done the same on another help with way less charted sounds), but if you think that's just too much of a complication, maybe we could use more general approximations. イヴァンスクルージ九十八_（会話） 07:24, 25 July 2018 (UTC)

Dialects of Lombard

It'd be wise to cover as few Lombard dialects (preferrably one, two or maximum three) as possible in this guide. Pick the most prestigious and/or most spoken ones and stick to them. Other dialects can be transcribed using the simple IPA template (or the IPA-all one), should there ever be a reason to provide multiple Lombard transcriptions in one article (generally, Wiktionary is a much better place for that). Kbb2 (ex. Mr KEBAB) (talk) 18:23, 24 July 2018 (UTC)

Of course, if the differences among all (or even most) dialects aren't that big, ignore my message. Kbb2 (ex. Mr KEBAB) (talk) 18:27, 24 July 2018 (UTC)

I see what you mean, but believe me, I couldn't be more general than this: dialects of Lombard can be really different from each other, and I tried to provide only the most widespread range of realizations (one or two per area). Moreover, if I took, for example, only the dialects of Milan and Bergamo, phonemes such as /dz/ and /ts/ would completely disappear despite being common in northwestern Lombardy and Switzerland. イヴァンスクルージ九十八_（会話） 19:02, 24 July 2018 (UTC)

Also, usually we mostly have to provide local pronunciations and choosing only a couple of dialects over the others wouldn't be very productive. I was also thinking of a similar expansion of Help:IPA/Neapolitan for the same reasons. イヴァンスクルージ九十八_（会話） 19:06, 24 July 2018 (UTC)

@IvanScrooge98: Then perhaps a better solution would be to provide a mere list of symbols (as on Help:IPA/Alemannic German) without diving into dialectal differences. Remember that the more complex the guide the less helpful it can be to laymen (it doesn't have to be unhelpful, but the more straightforward and non-complex the presentation the better). Kbb2 (ex. Mr KEBAB) (talk) 19:09, 24 July 2018 (UTC)

@Kbb2: Yes, at first I had this idea to make differences between EL and WL clearer at a glance, which still works for the consonant section, and because certain sounds are completely missing in some dialects while being basic in others. In other words, it's something different from "chose in Southern France is [ʃɔz] with that same sound you find in bosse in Parisian French". But maybe it wasn't such a brilliant thought. イヴァンスクルージ九十八_（会話） 19:26, 24 July 2018 (UTC)

In any case, I'm not sure if we should change this because there is not such thing as a standard Lombard pronunciation, thus dialectal comparison is still much more useful than for other languages. イヴァンスクルージ九十八_（会話） 07:48, 25 July 2018 (UTC)

@IvanScrooge98: I though so, but there are dialects that are spoken by more people than others. Kbb2 (ex. Mr KEBAB) (talk) 10:38, 25 July 2018 (UTC)

@Kbb2: you are right, but I personally wouldn't be able to say with certainty in which specific dialects certain things occur or not, nor to find references for that. And as I said, we can't let certain sounds aside just because they don't appear in more 'prestigious' accents. イヴァンスクルージ九十八_（会話） 11:20, 25 July 2018 (UTC)

@IvanScrooge98: Alright, so maybe we should have a table which only lists IPA symbols in an alphabetical manner (one column, no differentiation between Eastern and Western Lombard) and then provide two example words, one from each variety (Eastern and Western) without diving into further differences (if that's necessary, let's do it in a note). What do you say? As I said, that's precisely what we do on Help:IPA/Alemannic German. Kbb2 (ex. Mr KEBAB) (talk) 11:31, 25 July 2018 (UTC)

@Kbb2: I don't know much about Alemannic German, so I can't have a precise idea about whether the variations within its dialectal span can be compared to those of Lombard, but as long as we provide at least one example per macrodialect (Western vs Eastern), I think we might give it a try, if it makes things easier. イヴァンスクルージ九十八_（会話） 11:45, 25 July 2018 (UTC)

Alright, plain alphabetical order established. Regarding examples, I simply moved the ones we had before to their current place: if you think some are just too many, feel free to remove them, along with its equivalent(s) in the other dialect. イヴァンスクルージ九十八_（会話） 16:03, 12 August 2018 (UTC)

@IvanScrooge98: Thanks, this is better. Though we could work on reducing the amount of example words... Kbb2 (ex. Mr KEBAB) (talk) 17:50, 12 August 2018 (UTC)

@Kbb2: regarding this, I tried the same with charts for Emilian-Romagnol, but the result was not, ahem, much better. Do you think I should avoid that for those too? (oh, I used examples from only one dialect there.) イヴァンスクルージ九十八_（会話） 08:09, 24 August 2018 (UTC)

@IvanScrooge98: You should always avoid creating more than one row when you don't list specific dialects. It's meaningless. Kbb2 (ex. Mr KEBAB) (talk) 09:18, 24 August 2018 (UTC)

Columns too. — Ƶ§œš¹ _{[lɛts b̥iː pʰəˈlaɪˀt]} 15:29, 24 August 2018 (UTC)

Allophones

Hi @IvanScrooge98, as you can see in the page history, I replaced [ɑ] with /a/ in the text, since - as you already know for sure - it is just an allophone existing in some varieties to realise /a:/ or even /an/, but it can be found rounded too.

This considerations could be done for [œ] as well, being just an allophone of short /ø/ in some varieties; plus, this could be applied to the nasalised vowel, as they are allophones of /an/, /en/, /in/, /un/...

I think this philosophy would be better, as it allows to make more comprehensive phonemic general transcriptions; for more detailed transcriptions (like in local names or phonologic analysis), as they can contain an infinity of sounds within Lombard dialects, the link would be better addressed to the general IPA help page.

Please let me know your thoughts about this.

Thanks Stévan (talk) 13:00, 4 November 2022 (UTC)

@Stévan: I had missed the previous edit which had changed my ⟨ãː⟩ to ⟨ɑ̃ː⟩, so I’m obviously ok with that (I don’t even think there is a dialect with that realization). The thing is, these helps are meant to provide actual phonetic explanations for readers, so oversimplifying to the supradialectal archiphonemes might just not do the work. So while it would make sense to use ⟨aː⟩ and ⟨ø⟩ instead of ⟨ɑː⟩ and ⟨œ⟩ and at most add a note explaining there is some variation, I think we should retain the nasals because they bear a phonemic contrast in most WL ([san] "they know" vs [sãː] "healthy"). 〜イヴァンスクルージ九十八［IvanScrooge98］（会話） 08:18, 5 November 2022 (UTC)

@IvanScrooge98, I think there must be for sure some [ɑ̃ː] in the large Western area where "dark" <a> is realised before nasal (I recall some Sanga's work); but yes, that is allophone with [ɑŋ].

That's probably the point about nasalised vowels: not allophones of /an/, /en/, /in/, /un/, /yn/ - as I incorrectly reported in the last comment - but they're of /aŋ/, /eŋ/, /iŋ/, /uŋ/, /yŋ/: so, going back to your example, we still got phonemic contrast between /san/ "they know" and /saŋ/ "healthy" (there'd more to talk about how to transcribe those two, considering what happens when these words are followed by a vowel-starting one, but this is not the place).

I don't want to oversimplify, but [saŋ] and [sãː] have definitely the same phonemes, and the use of the first model allows to include more often Eastern pronunciations within the same transcription, e.g. /mi'la(ŋ)/ for "Milan".

Stévan (talk) 10:43, 5 November 2022 (UTC)

@Stévan: surely a supradialectal transcription could be implemented, but again, this is basically only done with English IPA here on Wikipedia (which is the reason {{IPAc-en}} is the only IPA template displaying in slashes rather than square brackets), and there is also another important point: {{IPA-lmo}} is mostly used for names of towns and toponyms in general, cases where the local pronunciation is clearly more relevant than others (not to mention sometimes the phoneme itself changes, e.g. the name for Bergamo in my WL dialect has a /u/ phoneme in the second syllable instead of the Bergamasque /e/ → [ˈbɛrɡum] and not [ˈbɛrɡɛm]); so it would be a bit tricky for an average reader to have to convert a transcription like /aŋ/ in /miˈla(ŋ)/ into the actual Milanese realization, especially considering they would generally have no clue whether it tends to [ãː], [ɑ̃ː] etc. or to [a], [ɑ] and so on. So actually a help for non standardized languages such as Lombard, where phonetic variation is pretty large, is already simple enough like this, by the way I see it. 〜イヴァンスクルージ九十八［IvanScrooge98］（会話） 11:55, 5 November 2022 (UTC)

@IvanScrooge98 Wait, no: I'm absolutely not talking about making a supradialectal transcription...

I'm just suggesting to use phonemic transcriptions generally, rather than phonetic ones, that in some cases could allow to use just one (as we said for /mi'la(ŋ)/), but not in others (like /ˈbɛrɡum/ and /ˈbɛrɡɛm/); but in the related pages you can just write /mi'laŋ/ and /ˈbɛrɡɛm/, since that's the phonemic transcription of local pronunciation.

An example can be the name of the town "Campiùn", which would be /kam'pjuŋ/ instead of [kɐm'pjʊŋ].

If we are talking about phonemes, then I see no reason to consider [ãː] as something different to /aŋ/, even because they can be found overlapped on the same dialectal areas, depending on the single speaker.

On the contrary, if this help page is for phonetic transcriptions, as I assume now when you mention squared brackets, then many and many symbols are missed here, maybe more than the existing ones.

--Stévan (talk) 14:51, 5 November 2022 (UTC)

@Stévan: that’s what I’m saying. Since we are using a simplified phonetic transcription (as is the case, again, for basically all of these pages), we are already cutting a number of realizations out (centering on a few varieties as the reference) in order to avoid cluttering this page with IPA symbols. So we either change this to a purely phonemic guide (and in that case we should also change the related template {{IPA-lmo}} so that it displays transcriptions inside slashes – but I guess that would need a more thorough and general discussion at Template talk:IPA-all), or we stick to this and list the possible realizations in notes. 〜イヴァンスクルージ九十八［IvanScrooge98］（会話） 17:35, 5 November 2022 (UTC)

Sorry @IvanScrooge98, I lost your last ping.

I see what you're saying, the current solution is a sort of hybrid between phonemic and phonetic; but it's actually closer to a purely phonemic one, considering that we're discussing just on nasalised and that we removed [ɑ] and [œ] for being allophones, so I think it's not a big deal to come to a substantially phonemic transcription, indicating all possibile allophones in notes or in dedicating a column within the tables.

I understand now the problem related to square brackets as well, but it should not be a problem if it would still get out with them instead of slashes (as even many English dictionaries make it the same); what matters IMHO is just that the template is used with symbols listed in this help page, and that very detailed transcriptions are addressed to general Help:IPA (even though I don't think we got so many "exotic" pronunciations reported here on en.wiki, to be honest). Stévan (talk) 18:39, 8 November 2022 (UTC)

@Stévan: I still find a phonemic transcription would ditch the whole purpose of a help, per the point I explained two replies back. Keep in mind that if there is need for a “help” it is because a lot of readers may not be familiar with how IPA works, and the near totality of them is also not familiar with this language. 〜イヴァンスクルージ九十八［IvanScrooge98］（会話） 07:39, 9 November 2022 (UTC)

@IvanScrooge98, I can't follow you throughout this last reply: the current page is already phonemic for every single IPA symbol (that's why we cut [ɑ] and [œ], because allophones of another phoneme), the one thing left out this logic are just the nasalised; so, saying that "phonemic would ditch the purpose" is equal to state that the current page is not making the job, except for the nasalised.

The fact that readers are not familiar with this language, and with the IPA, sounds to me like a further reason to avoid using a detailed transcription like [ãː], because: (1) it doesn't exist in English, and (2) [aŋ] does exists both in English and Lombard, being the actual phoneme.

I'm not saying better to not use [ãː] just in this help page (it would not make sense), I'm saying it should not be used with the template IPA/lmo at all; even because, if you grab 10 Milanese speakers, you would probably find that 3 pronounce [mi'lãː] and 7 [mi'laŋ], with none of them being able to distinguish the two forms, so it is just useless. Stévan (talk) 11:01, 9 November 2022 (UTC)

@Stévan: first of all no, the current page is not phonemic because we wouldn’t use ⟨ʃ⟩ when a dialect uses [h]. And my point is a too strictly phonemic approach risks to fail the purpose of this kind of help page, which is to help readers actually pronounce the words. In this particular case doing so would risk to lever the major dialectal variations to the point where a reader would have to reconstruct the local pronunciation (or the closest we can get) from scratch while being unfamiliar with it (so it is okay to use ⟨aˑ⟩ for minor variations thereof). I think we can cautiously convert nasal vowels to vowel+⟨ŋ⟩, but again we would need a note making a number of things clearer – i.e. variation between [ãː] and [ɑ] in Milan can hardly be treated the same as the [aː]–[ɑː] range. 〜イヴァンスクルージ九十八［IvanScrooge98］（会話） 11:31, 9 November 2022 (UTC)

@IvanScrooge98, I think we are not getting along on terms: /h/ is not an allophone of /s/ in Eastern dialects which use it, cause they have both and sometimes being distinctive, so there's no reason to join them (as we are neither talking about ⟨graphemes⟩); possibly, is /h/ to have some allophones, like [x], which obviously do not need to be mentioned here.

Instead, [aŋ]-[ãː] variation is definitely the same as [aː]–[ɑː], since [ãː] is exclusively a realisation of /aŋ/, and nothing else, even because it's very hardly distinguishable by speakers; there's nothing to be cautiously about in this case, it's a really plane case of allophones. Stévan (talk) 13:09, 9 November 2022 (UTC)

It is new to me that there are dialects that contrast /s/ and /h/, I’d genuinely love to learn more; but as you can see, for the way I conceived this page, [h] is one of the major allophones of the diaphonemes /ʃ/ and /s/ (as well as finally-devoiced /ʒ/ and /z/), and is listed for instance at Brescia. The /aŋ/–/ãː/ variation in stressed syllable tends to involve a change in length in dialects with generalized length contrast in stressed syllables (I think it would be confusing and/or inconsistent for a reader to see [aŋ] transcribed as short when in fact it is often long in dialects that preserve long oral vowels), and V+nasal clusters merge with plain V in certain instances of certain dialects. This is why it is a bit tricky. 〜イヴァンスクルージ九十八［IvanScrooge98］（会話） 00:07, 10 November 2022 (UTC)

@IvanScrooge98, I'm exactly talking about Brescia area: varieties using [h] have [s] as well, and not overlapped, so that e.g. "grasie" is never /'grahje/; long time ago I read about a few contrasts, but right now can't recall where (maybe Sanga, again), and that's why I'm ok with not kicking it off the table.

Going back to the main, for sure there is a contrast between /aŋ/ and /a:ŋ/ in Western varieties, as you can see in the Swiss Vocabulary - which is definitely phonologic - in words ending in ⟨-aan⟩ rather than ⟨-án⟩; in "nasalised speakers" they likely converge to [ãː], so - for their case - it's not a big deal representing it in one manner or the other.

Please make me an example of V+nasal clusters, cause I'm not getting the point, even though it looks transcribable just by cutting /ŋ/ off.

Last but not least, let me say I'm pretty sceptical about shortened nasalised vowels in unstressed syllables, as I tried to pronounce it many times, but it sounds like something totally far to any current speaker. Stévan (talk) 12:17, 10 November 2022 (UTC)

@Stévan: the fact that both phones exist does not mean they are not in complementary distribution (this [s]/[h] thing is mentioned at Eastern Lombard dialect and by the looks of it the two sounds are in free variation in the presence of /j/ – which however becomes [tʃ] when /s/ is debuccalized); anyway, be that the case or not, a reader should be able to tell the pronunciation without having to infer it from a transcription that is too broad; and by that not only do I mean that we should only list local pronunciations at Bergamo, Brescia, Milan etc. (so not /miˈla(ŋ)/ because [miˈla] is not a local form), but that in articles such as Lombard language we should avoid preferring a basically supradialectal /lomˈbart, luŋˈbart/ (or anything along those lines) to a list of the main variants. Ideally, we should never see /V(ŋ)/ with because a reader should see more clearly what output the phonemes actually combine into: yes to [miˈlaːŋ] if that is how we choose to transcribe it, but not to /miˈla(ŋ)/.

As far as short nasal vowels are concerned, I feel that is how I normally utter them in unstressed position.

(Be careful what Unicode characters you use, by the way: not ⟨g ' :⟩, but ⟨ɡ ˈ ː⟩).〜イヴァンスクルージ九十八［IvanScrooge98］（会話） 20:59, 10 November 2022 (UTC)

@IvanScrooge98, I 100% agree on using just the local transcription in pages of cities, municipalities, localities, etc... The use I was meaning for such comprehensive transcription as /miˈla(ŋ)/ is mainly for technical purposes, like comparatives with other languages: e.g. "Lombard /miˈla(ŋ)/, Emilian /'mlɛn/" (I'm inventing), in this case just to explain what is the general absence of internal syncope and the preservation of A in Lombard varieties.

I don't think it is needed to represent [aːŋ] instead of [aŋ] for what is now transcribed as [ãː], because [aŋ] and [ãː] already have the same syllable length (like [ʌr] and [ɜ:] in English): this is why I say they are hardly distinguishable by speakers, as the one difference is whether the two sounds are separated or "fused" in one, but total length is same. In fact, Swiss Vocabulary reports the usual-city-name as "Milán" rather than "Milaan" ("aan" is used in the Vocabulary too, but not for this kind of words).

And this is also why I'm saying that short nasalised vowels sound weird at a Western ear: pronouncing [lũ'ba:rt] makes it very similar to a hypothetical [lu'ba:rt] ("lubaart", in Ticinese letters), since the first syllable length is not the same as in [luŋ'ba:rt], sounding like the consonant is just chopped off; and that's definitely more like a Eastern pronunciation (like "tep" for "temp").

In synthesis: my opinion is for using [miˈlaŋ] in Milan and [luŋˈba:rt, lomˈbart] in Lombard language (using the right Unicode characters). Stévan (talk) 13:28, 11 November 2022 (UTC)

@Stévan: I must have misinterpreted what you meant then, I apologize; it does make sense actually. But the help page is going to need quite a makeover: I will proceed myself to implement the changes both here and in the related transcriptions, in due time.

(I invite you again to take a closer look at the Unicode symbols you use – ⟨ɡ ː ˈ⟩ – for any future IPA you may add on the project.) ;) 〜イヴァンスクルージ九十八［IvanScrooge98］（会話） 15:53, 11 November 2022 (UTC)

No worries, what matters is getting to the point.

Let me know in case you need something while doing it. Stévan (talk) 18:14, 11 November 2022 (UTC)