User talk:Kwamikagami

Your comments may be archived here after 48hrs

Word/quotation of the moment:

Astrology has no effect on reality, so why should reality have any effect on astrology? – J.S. Stenzel, commenting on astrological planets that astrologers acknowledge don't really exist

(Previous quotes)

The official state rainbow flag of Russia (official in JAO since 1996) Do you think the liberals are using these school shootings to further their anti-tragedy agenda? — Col. Erran Morad, Who Is America?, s01e01 yod-dropper — (when you need something that sounds like an insult)[1] ALL keys matter — response to the scale-wandering rendition of the national anthem at CPAC 2021 The Lunatic-in-Charge becomes the Lunatic-at-Large Lame duck à l'orange (AKA canard à l'orange) It is a mortifying circumstance, which greatly perplexes many a painstaking philosopher, that nature often refuses to second his most profound and elaborate efforts; so that often after having invented one of the most ingenious and natural theories imaginable, she will have the perverseness to act directly in the teeth of his system, and flatly contradict his most favorite positions. This is a manifest and unmerited grievance, since it throws the censure of the vulgar and unlearned entirely upon the philosopher; whereas the fault is not to be ascribed to his theory, which is unquestionably correct, but to the waywardness of Dame Nature, who, with the proverbial fickleness of her sex, is continually indulging in coquetries and caprices, and seems really to take pleasure in violating all philosophic rules, and jilting the most learned and indefatigable of her adorers. [...] The philosophers took this in very ill part, and it is thought they would never have pardoned the slight and affront which they conceived put upon them by the world had not a good-natured professor kindly officiated as a mediator between the parties, and effected a reconciliation. Finding the world would not accommodate itself to the theory, he wisely determined to accommodate the theory to the world. — Washington Irving, Knickerbocker's History of New York Pela primeira vez na sua vida a morte soube o que era ter um cão no regaço. For the first time in her life, death knew what it felt like to have a dog in her lap. — José Saramago, Death with Interruptions / Death at Intervals It is now generally accepted that the megaliths that make up Stonehenge were moved by human effort. — as opposed to by what? Anybody who says you only have yourself to blame is just not very good at blaming other people. — It's Happy Bunny When poppies pull themselves up from their roots and start out, one after the other, toward the sunset – don't follow them. — Slavko Janevski, 'Silence' And the dough-headed took their acid fermentation for a soul, the stabbing of meat for history, the means of postponing their decay for civilization. — Stanislaw Lem, Return from the Stars The Church says that the Earth is Flat, but I know that it is Round, for I have seen its Shadow on the Moon, and I have more Faith in a Shadow than in the Church. — (commonly misattributed to Magellan) In the early years of the study there were more than 200 speakers of the dialect, including one parrot. — from the WP article Nancy Dorian Mikebrown is unusually eccentric and not very bright. [...] Astronomers have not noticed any outbursts by Mikebrown. — from the WP article 11714 Mikebrown Ecce Mono Keep Redskins White! "homosapiens are people, too!!" a sprig of spaghetti "I've always had a horror of husbands-in-law." awkwardnessful anti–zombie-fungus fungus "Only an evil person would eat baby soup." (said in all sincerity)

Boma dialect

@Kwamikagami: Hirayama (1966)'s 琉球方言の総合的研究 talks about the prosody of the Tokunoshima Boma (母間) dialect on page 151-152. However, I am fine if you leave the redirect as redirecting only to Teke language. Chuterix (talk) 22:05, 17 July 2024 (UTC)

Yeah, there's nothing about that on the Toku-no-Shima page, and it's not in the list of dialects, so people won't know why a link directs them there. — kwami (talk) 22:10, 17 July 2024 (UTC)

Vocabulary comparisons

Hey Kwami, first of all, thanks for putting in the effort to make sure that Wikipedia articles stay encyclopedic, clean, and tidy. However, I strongly disagree that lexical comparisons in Papuan language articles are unencyclopedic. If you look at Bill Foley's chapter on Papuan languages of the Sepik region and many other sources, vocabulary comparison tables containing no more than 30 words are included because so little is known about those langauges. Oftentimes, basic vocabulary (anywhere from 20-100 words) and pronouns are all that we know about these languages, with very little grammatical information available.

These vocabulary tables strongly complement Ross's pronoun tables.

I know that Wikipedia is not Wiktionary - which was why I included no more than 30 words - but sample words for the world's least documented, most obscure language families are very important. Foley (2018) included tables comparing about 20-30 words, but his chapter was still encyclopedic and didn't get turned into a dictionary.

Such lists would not be very suitable for major languages such as German or Luxembourgish, but they are absolutely essential for Papuan languages, which are undoubtedly the world's least known languages. Wikipedia articles about sparsely documented Amerindian languages also list the few words that are attested in those languages. It would be completely inappropriate to remove all those words on the basis of Wikipedia's NODICT guideline, which is a guideline, not a set-in-stone policy, meant to discourage the creation of dictionary-type stubs, rather than to dissuade editors from including useful vocabulary comparison tables in linguistics articles.

And if you really want to delete these painstakingly compiled and highly useful tables, please move them to Wiktionary appendices first.

Pinging @Austronesier: and @Womtelo: for opinions. Please consider supporting to keep these vocabulary tables, as they are of utmost importance to historical linguists. For now, I am restoring the tables, since I am certain that I have set up a strong case for why these tables should be kept. — Sagotreespirit (^talk) 22:12, 23 July 2024 (UTC)

There is a distinction between brief encyclopedic basic vocabulary tables and lengthy dictionary-type vocabulary tables that do not belong in encyclopedic entries. Please consult language survey articles and chapters to see what I mean. Tables comparing up to about 30 non-cognate words are encyclopedic and typically do not need to be shuffled into appendices. In language survey papers, articles, and books, non-encyclopedic tables such as Swadesh lists with 100-200 words usually get moved into appendices. The vocabulary tables that I have included in Papuan language family articles are completely encyclopedic. Thus, there is no reason to remove them without proper discussion. Since this is contested by multiple editors including both me and Womtelo, I strongly do not recommend removing any of the tables for now. — Sagotreespirit (^talk) 22:44, 23 July 2024 (UTC)

Thanks @Sagotreespirit. I agree with your arguments here (and on my Talk page). The tables in question represent abundant efforts by earlier contributors (not necessarily by myself), and I find it a bit blunt to delete so much data so fast, without any discussion or search for consensus. I'm talking about both East Papuan and Baining languages, to cite the first two I noticed; but I now see there are many more pages that Kwami had unilaterally decided to empty of their lexical content; I find that unreasonable.

In fact, I think it's all the more useful to keep such lexical lists, that precisely the small language families in question are still being actively researched. As you (Sago) said, we wouldn't need such tables under entries on German or Dutch; but researchers & students will find it very useful to have easy access to well-structured lexical lists on WP (as long as the data itself is sourced and not OR); this can only help science grow, much easier than if people were forced to retrieve every data point individually.

Thanks also to @Kwami for constructively taking part in the discussion. Best -- Womtelo (talk) 23:16, 23 July 2024 (UTC).

If the point is to use WP as a repository for raw data, there are other places to do that -- Wiktionary appendices, Wikibooks, etc. That's not the point of an encyclopedia, so all of these lists are unencyclopedic. Doesn't matter how well referenced they are, that's not the purpose of WP. Foley may do that in his publications, but those serve a different purpose.

I don't have a problem with such vocab lists in the individual language articles (what you call 'brief encyclopedic basic vocabulary tables'), and if you moved all the data there I wouldn't have a problem with it. But what does it have to do with the family? For the average reader, nothing.

If they're cognate sets, sure. I left those in. Either just the reconstructed forms, or better yet those forms with reflexes in a sampling of languages, along with the same for phonemes. That would tell the reader something. It's also okay IMO to list cognate sets without a reconstruction, if we have a RS they are cognate but the reconstruction hasn't been done. There were a few cases of that, and I left those in too.

What's the point of putting the lists for various languages together? It can only be for comparison. But what are we comparing? There's no explanation, and we shouldn't expect the reader to do the historical comparative work. Also, how is the reader supposed to know which words are cognate, and which are not, when they're all scrambled together? Some seem obvious, but sometimes even obvious resemblances turn out to be coincidental (e.g. English much and Spanish mucho). That occurs in New Guinea as well: e.g. the Bunaq word for 'hand' is ton. If we placed that in a row with East Timor tana, Alor-Pantar tan and Kalamang tan, the reader would naturally assume that it's related, but it's a false cognate. It took Usher years to figure that out, and it would be entirely unreasonable to expect the reader to figure that out just by looking at the table.

Sometimes words that are cognate don't look related (e.g. the Armenian word for 'three'), and the reader might therefore assume the words in each row are all cognate (otherwise why list them together?). Worse yet, sometimes these word lists include the reconstructed forms in the left column, which essentially tells the reader that they are reflexes of that form when in fact many are not. So these lists are useless at best to the reader, and sometimes highly misleading. Their only use is as raw data, and again that's not what an encyclopedia is for.

We could link to the original DB's in an external links section, or move these lists to Wikt or Wikibooks and link to them there. — kwami (talk) 23:40, 23 July 2024 (UTC)

As an analogy, in our astronomy articles we give orbital information. For some objects the orbits are highly uncertain. These vocab tables are the equivalent of giving the raw positional data of multiple astronomical observations, including some that might not even be of the correct object, and expecting the reader to work out the orbit from that. Instead, we rely on the experts to calculate the orbit, and if the raw data is relevant somehow, we can link to it. — kwami (talk) 00:03, 24 July 2024 (UTC)

Also, Sago, you ask me to do your work for you. If you want to make this information available, that's a task you've taken on for yourself, and it's up to you to carry it out. I might help you if I saw the value in it, but that is irrelevant when it comes to deleting unencyclopedic content. You can always go back into the article history and recover it, so deletion does not need to wait for it to be transferred. — kwami (talk) 00:21, 24 July 2024 (UTC)

Thanks for your answers Kwami, but I'm afraid I disagree with your arguments. If you tell me that a family called XXX exists and has 15 members, that is very little information and is telling us nothing. But if you give me a list of 20 words in these 15 members, now that’s what many readers are looking for. Anyone interested in the existence of a family will also want to get a good idea, at a glance right there on the page of the family, of how internally diverse that family is. That is extremely useful information that is telling so much more on that subgroup, than just a name and a list of languages. Some families are terribly homogenous like Nuclear Polynesian, while others are surprisingly heterogenous; and just a lexical list can already tell us so much.

People sufficiently curious to get interested in Baining or Leonhard Schultze languages will be smart enough to know what you're saying here about false cognates and the rest. And if they mistake a list of translation equivalents for cognate sets, so what? Is that reason enough to deprive the rest of us of the valuable data? We can easily add a disclaimer template that will remind people that list of translation equivalents are not cognate sets, and voilà the job's done. But it is foolish to delete so much useful information, which target readers (students, researchers) actually need badly, just because some imaginary reader might interpret them wrongly. I find that a very strange way of reasoning, really.

In sum, these painstakingly-created tables are very useful, and much welcome on pages of families -- much more so than on entries of individual languages. (Perhaps one thing that could be done, would be to make them collapsible; but I don't think that's necessary.) -- Womtelo (talk) 00:12, 24 July 2024 (UTC).

That's the purpose of cognate sets. As I said, I agree those are encyclopedic. But random sets of words that we falsely imply are related is misinformation and do not belong. If they're useful to the historical linguist, fine, but put them where they belong.

And they don't tell us anything about the family, and they do not show us how diverse the family is. For that they would have to be cognate sets. I could create a list of random words that makes the Romance languages look like dialects of a single language, or like a collection of unrelated isolates. Neither would serve any useful purpose. In fact, Serbo-Croatian and Hindustani speakers do this all the time to 'prove' that their languages are either identical or completely unrelated. — kwami (talk) 00:19, 24 July 2024 (UTC)

No, you're not getting it. If you only accept a list of cognate sets, then you're forcing a view of unity in the family, which is not necessarily a fair representation of its diversity; you're doing precisely what your Hindustani and Serbo-Croatian speakers do all the time, namely cherry-pick examples to suit an agenda. What is needed is precisely the opposite: random lists of words (à la Swadesh lists), absolutely avoiding the criterion that they should be cognate sets. What we need are simply lists of translation equivalents within that family, whether or not they're mutually cognate.

So in Romance you'd have aller, andare, ir for 'go'; bras, braccio, brazo for 'arm'; femme, donna, mujer for 'woman', etc. No-one in their right mind will be so foolish as to believe these are cognate sets! Why do you assume that will be "falsely implied"? That's just wrong. (and there can always be a warning message/ disclaimer in small characters reminding people that the list is not of cognate sets.)

Such lists would provide an excellent overview of the internal homogeneity vs. heterogeneity of a lg family: both (1) diversity due to sound changes within cognate sets, and (2) diversity due to lexical replacement across cognate sets. If you only display cognate sets, then you can show #1 but not #2, and that becomes a cherry-picked sample biased towards homogeneity. For the purpose of showcasing a family's internal homo- / heterogeneity, it is thus crucial that there should not be any requirement of cognacy. (In addition, color coding or something could identify cognate sets, a bit like in this Romance table).

Ideally, the list of meanings would be constant across families (like Swadesh, but shorter; perhaps 15 or 20 meanings of basic vocab); but for some little-documented families we would provide whatever small lexical info there is. -- Womtelo (talk) 09:55, 24 July 2024 (UTC).

For better visibility to the community, I suggest to bring this up in WT:LANG. Kwami and also, @Womtelo and Sagotreespirit: as you already have had the opportunity to shape your thoughts in this discussion, you can each probably repeat them in a condensed form there (without reproducing the back-and-forth, at least to this point). My 2 cents in brief here. I have seen many tables that go much beyond 20 items. I find them way too massive, visually dominating and literally pushing our painstakingly-created prose (c'mon, what's painstaking about retyping wordlists, often to the point of producing blatant copyvios?) to invisbility. Frankly, I hate them ;) So ~15 items per table, and all tables headed by at least two or three lines of parsable prose with substantial content (NB: wordlists of languages not known to our readers are per se imparsable and only serve an "impressionistic" purpose). That's what I consider the absolute maximum for a tertiary source like Wikipedia, anything beyond that is not really helpful for our readers. FWIW, collapsed tables don't appear collapsed on many mobile browers, thus in the very case when navigatability matters most.

Tables don't have to be cognate tables, but for established language families and subgroups, cognate tables give a better feel of what these languages have in common; after all, these languages by their very nature are relevant to the topic of the article (= the language family) since they do share cognates with regular sound correspondences. In order to capture diversity, a cognate table can serve the purpose as well. E.g. for Austronesian languages you often see those silly "mata, mata, mata, mata, mata...lima, lima, lima, lima, lima"-tables which may lead naive readers to think that Austronesian languages are alike all over the oceans; ideally, we should present them also cognate sets like Paiwan sapuy, Palauan ngau, West Damar oso, and Kusaiean e :) Ok, not all that brief, but that's how I feel about it (= my editorial judgement). –Austronesier (talk) 15:57, 24 July 2024 (UTC)

Okay, lists up to 20 or so basic items, as long as we make it clear that they're not cognate sets. — kwami (talk) 20:09, 3 August 2024 (UTC)

The current lists are fine. Bill Foley uses 20-30 basic items, and that's not even enough to satisfy the rigorous standards of many historical linguists. I wouldn't go up to 40 (=ASJP). It's not like huge Swadesh lists are completely bloating up the articles. — Sagotreespirit (^talk) 03:06, 15 August 2024 (UTC)

And if you really want to remove the tables, please move to Wiktionary first and link to them. — Sagotreespirit (^talk) 03:06, 15 August 2024 (UTC)

WikiProject

Hi, I see you've contributed a lot to Hadza people, would you be interested in a taskforce on oral tradition? Kowal2701 (talk) 19:44, 3 August 2024 (UTC)

Thanks, but probably not. — kwami (talk) 19:54, 3 August 2024 (UTC)

No worries Kowal2701 (talk) 19:57, 3 August 2024 (UTC)

Question mark

No the question mark is not needed here. What makes you insist that it is? Fdom5997 (talk) 03:11, 4 August 2024 (UTC)

Our source that it's not clear if it's a distinct phoneme. Parentheses only mean that it's marginal or found in loans. — kwami (talk) 03:24, 4 August 2024 (UTC)

Check the reliability

Hi. You have recently removed a para in the page: 'Telugu Language', stating that it is not a reliable source. I am sharing an alternate source and I request you to check if this can be considered. If yes, I shall revert the edit by adding this as the citation.

Source: (Page no.4 from)

https://rmrl.in/wp-content/plugins/book_lists/admin/pdf_books/38%20-%20Harappan%20Heritage%20of%20Andhra_%20A%20New%20Interpretation%20(2010).pdf

Waiting for your reply. Thank you. Anandwiki.ind (talk) 21:12, 7 August 2024 (UTC)

Not peer-reviewed. Given that it appears to be speculation built on sand (we don't even know if the harappan 'script' was a writing system, let alone that it was in a Dravidian language), I'd want something better per FRINGE and WP:RS. — kwami (talk) 21:23, 7 August 2024 (UTC)

The point is not about whether Harappan script was Dravidian origin or not.

The source only tells that Mahendra Pallava did lay few words of Telugu in cave inscriptions and that Iravatham Mahadevan used them in his study, or 'speculation' or whatever. I'm not making any statement regarding the Indus script. Anandwiki.ind (talk) 21:26, 7 August 2024 (UTC)

Yes, but it's still FRINGE material and not a RS. You could try using their source directly. — kwami (talk) 21:27, 7 August 2024 (UTC)

What do you mean by "their source directly"? Anandwiki.ind (talk) 21:29, 7 August 2024 (UTC)

Cite the source that they use. I'm assuming that the inscriptions and their analysis as Telugu have been published somewhere. — kwami (talk) 21:31, 7 August 2024 (UTC)

Fine. I shall use the sources which include:

1. The original data of the inscriptions

2. The original data of the word nāgabu (but this is in non-English language)

Will it be fine? Anandwiki.ind (talk) 21:33, 7 August 2024 (UTC)

Sources don't need to be in English, they just need to be from a reliable source (e.g. a secondary reference used by paleographers), not a website or someone's personal research. — kwami (talk) 21:40, 7 August 2024 (UTC)

I wouldn't draw any conclusion that "Harappan was proved to be Dravidian based on these words".

Rather, the presence of Telugu words as the titles for Mahendra Pallava in his cave inscriptions were some of the important things in Telugu history. That's the only point I would mention. Anandwiki.ind (talk) 21:28, 7 August 2024 (UTC)

Sure. But you need a reliable source. — kwami (talk) 21:32, 7 August 2024 (UTC)

Wiktionary

If you are insistent on removing Papuan vocabulary tables, please move them to Wiktionary first. Please don't completely remove useful comparative tables just because you think they look out of place. Plenty of people (me, Womtelo, the NGW guys, and many others) find them to be incredibly useful, even if you don't. At least insert Wiktionary links to transwikied tables if you really want Wikipedia articles to cater solely to syntacticians and be nearly completely devoid of useful lexical data. — Sagotreespirit (^talk) 02:50, 15 August 2024 (UTC)

I'm not going to do your work for you. It's up to you to preserve unencyclopedic information elsewhere. Anything you added here is preserved in the page history and can easily be recovered.

You gave criteria for inclusion, which were already stretching appropriate content for an encyclopedia, but then insisted on violating your own conditions. As others have told you, lists of random vocab belong on wikt (perhaps as an appendix), not on WP. — kwami (talk) 03:42, 15 August 2024 (UTC)

I did not violate my own conditions. Your decisions of exactly which tables to exclude were arbitrary and not sufficiently explained. Nevertheless, I will copy everything to Wiktionary and will provide interwiki links and templates from Wikipedia, because this is just going on forever. If you want to remove any more comparative tables, then please let me know what you have removed. It is very time-consuming to compile these tables, and even more so to constantly have to restore them and make sure that no errors have crept in due to your constant trimming. Please at least try to be more considerate like all of the other users by carefully discussing your concerns first, instead of repeatedly deleting entire sections across all Papuan language family articles without properly notifying anyone. In any case, I sincerely appreciate your prolific contributions to Wikipedia. — Sagotreespirit (^talk) 04:52, 15 August 2024 (UTC)

You said 15 to 20 lexical items, not 40 to 60. It's supposed to be a sample, not an exhaustive vocabulary. Wiktionary is where you post vocabularies.

You have been properly notified, and you have no consensus to turn Wikipedia into a dictionary. — kwami (talk) 05:06, 15 August 2024 (UTC)

None of the tables have 40 to 60 items. They all have 30 items maximum. These are not exhaustive vocabularies. I know what's supposed to be encyclopedic and what's supposed to be dictionary-length. The Languages and Linguistics of the New Guinea Area (2018) and many other encyclopedic handbooks have plenty of word lists of approximately 30 items in their chapters. Anything longer than the word lists in the ASJP or Global Lexicostatistical Database would be dictionary-like, and I'm not including any of those in Papuan language family articles.

I am not turning Wikipedia articles into dictionary entries. I am balancing out phylogenetic trees, classification histories, geographical distributions, and everything else with vocabulary tables. This is fully encyclopedic. Wikipedia articles don't have to be restricted to dense grammars with completely no word lists in order to be encyclopedic. — Sagotreespirit (^talk) 05:25, 15 August 2024 (UTC)

So you're saying you're pushing for only twice what you agreed to. But they go up to 50, and completely dominate the articles. That's exactly what you said you were not trying to do.

You've brought this up for discussion, and did not get consensus, but still insist that you're in the right because that's what you want.

Your strawman arguments ("with completely no word lists") show that you're arguing in bad faith. I'm not going to waste my time with someone who isn't willing to collaborate. — kwami (talk) 05:32, 15 August 2024 (UTC)

Look, I am collaborating and cooperating now by agreeing to copy the tables to Wiktionary. I don't need to add any more tables, and you don't need to delete any more tables. We're both fine with this status quo. Moving on. — Sagotreespirit (^talk) 05:38, 15 August 2024 (UTC)

Correct. I don't need to delete any more tables. I already removed the ones that violate the conditions you agreed to and overwhelm the text. Now that you're copying the info to Wikt, you shouldn't need to continue to edit-war over restoring them here.

BTW, if you're dissatisfied the results of the discussion, you're welcome to open it again. Or ask Austronesier their opinion. They might be more tolerant than I am. — kwami (talk) 05:50, 15 August 2024 (UTC)

@Sagotreespirit Just FYI, Wiktionary is not a dumping ground for things that have been removed from Wikipedia. I have no view on whether these should or shouldn't be included (on WP or Wikt), but they need to pass Wiktionary's criteria for inclusion, which means they need to be verifiable, and included in accordance with our entry layout; what we don't want are random tables of vocabulary. Theknightwho (talk) 16:54, 15 August 2024 (UTC)

@Theknightwho: Thanks, I am aware that Wiktionary lists should be properly formatted as appendices, and should preferably be standardized Swadesh lists or their subsets so that they can be consistent with existing lists compiled by other Wiktionary users. The lists will have to be redone and not simply copied as-is from Wikipedia. — Sagotreespirit (^talk) 00:34, 16 August 2024 (UTC)