Talk:List of language families

Semitic and Persian

I miss any mention of languages of the Middle East, or a reference to an article about these families. Or maybe I´m missing something? — Preceding unsigned comment added by 80.203.105.96 (talk) 20:56, 29 October 2014 (UTC)[reply]

For Semitic, see under Afro-Asiatic, for Persian-Iranian under Indo-European. Eklir (talk) 21:23, 29 October 2014 (UTC)[reply]

Altaic

Is Altaic actually controversial? I know it's controversial to have it include Japanese and Korean, but I thought the relationship between Turkic, Mongolic, and Tungusic languages was well-established. Similarly, how is Afro-Asiatic controversial? I've read that there's some controversy over at least some of the Chadic languages, but is the relationship of Semitic, Egyptian, Berber, and Cushitic to one another disputed? At any rate, both families are listed in just about every general reference work I've ever seen. john k 08:33, 5 March 2007 (UTC)[reply]

"Altaic" even in the narrow sense is very far from being generally accepted. Numerous comparative linguists who work on Turkic, Mongolic and Tungusic languages do not accept the Altaic hypothesis or at least do not consider it proved. Recently there has been a lot of research on how lexical and typological correspondences between the "Altaic" languages can be explained as contact phenomena. My impression is actually that opposition to the Altaic theory has increased during recent decades. --AAikio 13:44, 6 March 2007 (UTC)[reply]

Fair enough, although I do think that our discussions of this tend to confuse the very controversial issue of whether Ainu, Japanese, and Korean are included with the less controversial issue of whether it is a group at all. The first is a not widely-accepted hypothesis. The second is a widely-accepted usage that has come into question in recent years. john k 14:57, 6 March 2007 (UTC)[reply]

These listings are particularly dubious given that we have the entirely hypothetical "Aegean languages" listed as though they are unproblematic, while we list problems with all kinds of widely accepted language families. john k 08:34, 5 March 2007 (UTC)[reply]

I have removed the highly speculative Aegean languages from the list. Nilo-Saharan and Khoisan are both less accepted than Afroasiatic (see Joseph Greenberg#African languages). And I share AAikios viewpoint that Altaic is indeed still quite controversial. ·Maunus· ·ƛ· 13:49, 6 March 2007 (UTC)[reply]

The map of languages picture

In the part which has India, I found a mistake. Where they labeled Marathi and Gujarati, they put them in the wrong places. If I am correct, they should be switched because currently Marathi is in Gujarat and Gujarati is in Maharashtra. —The preceding unsigned comment was added by 67.81.107.17 (talk) 14:38, 14 April 2007 (UTC).[reply]

What's going on?

Ethnologue, Ethnologue: Languages of the World, 15th ed. (2005) listed the following languges in order of speakers,

1. Indo-European 2.562 billion 44.78%
2. Sino-Tibetan 1.276 billion 22.28%
3. Niger-Congo 358 million 6.26%
4. Afro-Asiatic 340 million 5.93%
5. Austronesian 312 million 5.45%
6. Dravidian 222 million 3.87%
7. Altaic 145 million 2.53%
8. Japanese 123 million 2.16%
9. Austro-Asiatic 101 million 1.77%
10. Tai-Kadai 78 million 1.37% —Preceding unsigned comment added by 4.239.87.2 (talk) 02:35, 16 March 2008 (UTC)[reply]

Any objection to adding these percentages (or updated ones) to the article? --JWB (talk) 12:49, 5 October 2009 (UTC)[reply]

We should go with the 16th ed,[1] and also round off to 2 sig.fig.- I doubt they're as accurate as this implies.

IE - 46%
ST - 21%
NC - 6.4%
AA - 6.0%
AN - 5.9%
Drav - 3.7%

Maybe we should stop here, because then we get "Altaic" 2.3%, Japonic 2.1%, Korean 1.1%, but those could all be argued to be Altaic at 5.5%. Then there's MK at 1.7%, TK at 1.4%, etc. kwami (talk) 21:43, 5 October 2009 (UTC)[reply]

OK, I added whole percentages only, summing to 92%, and moved Japonic up to its rank in Ethnologue. Next would be Nilo-Saharan at 38m, less than 1% and less than half of Tai. --JWB (talk) 02:24, 6 October 2009 (UTC)[reply]

When you get down to low figures, the difference between 1 and 1.4 is huge, so we might want to stick to 2 significant figures? Or else round IE and ST off to 50 and 20%? kwami (talk) 05:46, 6 October 2009 (UTC)[reply]

The former is fine. The latter would add a second digit that is actually wrong. Defining accuracy in absolute terms is as legitimate as in relative terms, if the focus is the world population, but then I don't want people complaining that Altaic or Tai are being shortchanged either. --JWB (talk)

Commented out bibliography

The editor who put the bibliography in there does not understand lists. This is a list, not an article proper. Lists list items that are offloaded from some other article or are too big for the other articles. The explanation is taken care of in the other articles. In this case the bibliography was copied from Language family and it was tacked after the old end of this article, which ended with external links. There are no notes in this article as references are not necessary in a list. Some lists have them but this one has no notes. The blue internal links take care of the references. There is no point in duplicating this bibliography so I am commenting it out.

One of the ethnic groups.

please want to the of adansi, one of the ethnic`s group. thanksI--196.29.232.90 (talk) 19:59, 31 January 2011 (UTC)italic text[reply]

Sorry, I do not understand your question. Adansi appears to have something to do with the Akan languages — which fall under the Niger-Congo family, if that helps. --Trɔpʏliʊm • blah 21:58, 31 January 2011 (UTC)[reply]

Japonic

User:Trikemike is engaged in an edit war in violation of WP:BRD. He is claiming that the Japonic languages are automatically included in Altaic. He is wrong. This is a minority view even among Altaicists. --Taivo (talk) 10:29, 30 March 2011 (UTC)[reply]

User:Trikemike continues to edit without discussion and inserting his POV. Help? --Taivo (talk) 10:43, 30 March 2011 (UTC)[reply]

Japonic definitely needs its own entry as Altaic is not universally accepted, and doesn't even include Japonic according to most proposals.·Maunus·ƛ· 11:32, 30 March 2011 (UTC)[reply]

I would agree that Japonic needs separate inclusion. The inclusion of any form of "Altaic" in a list of "fairly often recognized families" seems suspicious, really. (Switching to just Turkic wouldn't even change the figures all that much.) --Trɔpʏliʊm • blah 18:47, 30 March 2011 (UTC)[reply]

I agree with you, especially since neither Mongolic nor Tungusic is a blip on the count. --Taivo (talk) 19:36, 30 March 2011 (UTC)[reply]

To clarify things, I've added a note that Altaic is Turkic, Mongolic, and Tungusic, but that the majority of speakers are Turkic. --Taivo (talk) 01:32, 31 March 2011 (UTC)[reply]

"Altaic"

"Altaic" is not universally accepted as a valid genetic unit. Indeed, it is only accepted by a minority of historical linguists as valid. Therefore, in Wikipedia, we nearly always include the word "controversial" to describe it so that casual readers do not get the idea that it is widely accepted by scholars. --Taivo (talk) 13:55, 1 April 2011 (UTC)[reply]

Correct.·Maunus·ƛ· 14:14, 1 April 2011 (UTC)[reply]

Uralic (finnic-ugrian) languages

Where do the uralic languages belong in these groups? — Preceding unsigned comment added by 68.39.110.172 (talk) 08:28, 11 November 2011 (UTC)[reply]

Uralic is a top level family by current knowledge and is listed in the Europe/Northern Asia section, AFAICS. --Trɔpʏliʊm • blah 16:59, 15 November 2011 (UTC)[reply]

Sorting nonstandard proposals

Some time ago I introduced a nested structure for this section, to illustrate a difference between the far-reaching a-continent-or-two proposals (Amerind, Borean) and narrower proposals that group two to half a dozen neighboring, perhaps shallow-ish families (like Altaic or Macro-Ge). However, I now wonder if this is a good idea: it gives the illusion of a "known taxonomy" where there is debate only on what groupings have been demonstrated in sufficient rigor, not on the macro-structure itself. There is a small disclaimer against this, but I wonder if it is sufficient. — Tropylium 01:11, 23 October 2014 (UTC) — continues after insertion below[reply]

You seem to refer to this edit. — Sebastian 20:53, 25 October 2014 (UTC)[reply]

It also seems unclear to me what should be done when one proposal groups A and B; another groups A, B and C; a third groups A, B, D and E. We could place AB under either of the larger proposals, but does this add up to implicit support for whichever is chosen?

On the other hand, I am unsure if there are other better ways to sort these. Alphabetic ordering seems unillustrative; strictly geographical sorting would run into many of the same problems. Perhaps a distinction between linguistically entirely speculative groupings (frequently based on anthropological considerations) and groupings for which actual morphological, phonological, etymological etc. reconstructions have been proposed might help.

Or perhaps these articles do not even belong on a list of known language families, and the topic of deep-level connections between language families should be a separate article entirely? --Trɔpʏliʊm • blah 01:11, 23 October 2014 (UTC)[reply]

It is great to have the list in a nested structure; alphabetic sorting is just the default solution if editors can't agree. The disclaimer is indeed a bit lifeless; it would be much more useful if it explained what the disagreements are. Better yet, why not mention the conflicts right next to the macrofamily entry? As in "* Pontic (Conflicts with Nostratic because language family X is claimed by both)" (Sorry, I had to guess; I didn't find mention of that conflict in either article.)

The ABCDE grouping question is a good one. How often does that happen? If in less than 20% of the entries, then it still can be handled with mentions next to the entries. If more often, then the question is: How important is it to chronicle all these differences? Realistically, we have to reconcile our desire to be fair with our limited resources. So I think it would be a healty compromise to write "This list follows So-And-So's proposal of 20xx" and provide references to alternative proposals, so readers can compare for themselves.

I'm not sure what potential benefit you see in a distinction between speculative and actual groupings; it seems to me that can only be achieved by adding yet another layer of complexity. Whether a grouping is speculative can be less obtrusively shown e.g. by adding an asterisk. — Sebastian 20:53, 25 October 2014 (UTC)[reply]

I suspect that there actually might not be much of a nested structure at all, beyond the half a dozen speculative megafamilies. If we take Indo-European as an example, it is included in the following proposals:

Proto-World (linguistically speculative), which covers everything.
Borean (linguistically speculative), which covers almost everything in Eurasia & then some.
Pontic: IE + Northwest Caucasian
Indo-Uralic: IE + Uralic
Nostratic: several views that are not mutually compatible
Eurasiatic: a couple of distinct views

We have PW ∋ Borean ∋ everything else. Pontic clashes with everything else (as well as with Ibero-Caucasian and North Caucasian). Indo-Uralic can be included in all versions of Nostratic and Eurasiatic (but it clashes with Ural-Altaic, Uralic-Yukaghir and Uralo-Siberian). Most versions of Eurasiatic fit under some version of Nostratic, but not by means all of them (and there is no real cutoff point for when a hypothesis stops being a version of "Nostratic" and starts being a version of "Eurasiatic"). --Trɔpʏliʊm • blah 21:32, 30 October 2014 (UTC)[reply]

Best place for the list of proposed language stocks

Tropylium suggested in the previous section, regarding the list at Proposed language stocks:

Or perhaps these articles do not even belong on a list of known language families, and the topic of deep-level connections between language families should be a separate article entirely?

This is a very good question; I feel that a better place would be in the article macrofamilies, which would also solve the problem Tropylium raised that the disclaimer may not be strong enough. That article is currently a stub, but it's clear that that's not for lack of interesting things to write about the topic, and a list would de-stub the article. If the name "macrofamily" doesn't suit everyone, we could rename that article to a more generic name such as "proposed language family". — Sebastian 20:53, 25 October 2014 (UTC)[reply]

We also have a reasonably detailed article Classification schemes for indigenous languages of the Americas. Perhaps similar articles could be written for Africa, Australia and Eurasia as well. --Trɔpʏliʊm • blah 21:48, 30 October 2014 (UTC)[reply]

Wide distribution but few speakers

Do we actually have any reason to believe that the number of people speaking Eskimo-Aleut, Na-Dene, Algic, Quechuan or Nilo-Saharan has ever been substantially greater than it is right now? The current phrasing makes it sound like that the numbers of people speaking these have recently plummeted for whatever reason — but I imagine their non-top-10 status has much more to do with the colossal growth in world population during the last 150 years, which has mainly involved peoples who do not speak languages of these families. --Trɔpʏliʊm • blah 12:46, 7 March 2015 (UTC)[reply]

Good point. (But then, I don't think they've ever been among the top 10. Nilo-Saharan may actually have, and have had for centuries, the most speakers of those, let's say many (four, at least?) millions in 1000 AD, and Quechuan a few (two, at least?) millions, but the rest not substantially more than 0.1 million each: let's say 0.2 million for Algic, 0.1 million for Na-Dene and 0.05 million for Eskimo-Aleut, to provide some ballpark/order-of-magnitude guesstimates.) I'm not sure how to rephrase it, however. --Florian Blaschke (talk) 21:25, 7 June 2015 (UTC)[reply]

Actually, the phrasing may not need to be tweaked as much as you might think. All of these language families have, indeed, declined in number of speakers over the last 200 years, especially the Native American ones. While they were never in the major category, their speaker numbers are infinitessimal compared to the era when they were actually spoken by hundreds of thousands, if not millions. (Eskimo-Aleut is probably not in that category, but Quechuan and Algic certainly are). At the time of the Conquest, Mexico City (née Tenochtitlan) was one of the largest cities in the world, easily larger than any in Europe. It is now understood that the diseases that arrived with Columbus wiped out perhaps as much as 90% of the population of the Americas before the first colonists arrived. Out of the Algic languages, three-quarters of them are either extinct or down to their last handful of speakers. --Taivo (talk) 23:23, 7 June 2015 (UTC)[reply]

That's probably true, but I don't think the actual point of that category is the population history at all, but rather the wide geographic distribution.

Also, the recent sortable format probably obviates the need for the remark at all. A simple re-sorting of the list by number of languages will reveal e.g. Nilo-Saharan having plenty of variety, or Algic, Na-Dene and Quechuan having more variety than the much larger Turkic or Uralic. --Trɔpʏliʊm • blah 13:11, 8 June 2015 (UTC)[reply]

Taivo: Hrm, yeah, that's true. So my guesstimates should be understood as conservative lower bounds. I don't have the literature, but serendipitously, I just came across this essay, which cites literature that gives more well-founded estimates for Mexico. 30 million in 1521 – so you are saying there were even several times more than that in 1492?

Tropylium: I don't think it is fair to say that Quechuan, a quite close-knit family, has more internal variety than Turkic, let alone Uralic. Even if language division is essentially comparably handled in all three cases, a family with as few as two or three distantly related members can have more variety than a family with dozens of closely related languages. (Note that Linguasphere has Quechuan as a "net", Turkic as a "set" and Uralic even as a "phylozone". The time-depth of Quechuan is given as 1300 years by Kaufman in Atlas of the World's Languages, that of Turkic usually as 2000–2500 years, I think, and that of Uralic as at least 4000 years.) --Florian Blaschke (talk) 01:24, 9 June 2015 (UTC)[reply]

Yes, Florian, there were probably at least twice that many residents of Mexico in 1492. The diseases that Columbus brought (primarily smallpox and measles) spread Indian to Indian like wildfire. When the first colonists arrived on the shores of North America in the first decades of the 17th century, they found a lush wilderness with very few people. There is no environmental reason for that lack of population--it is the aftereffect of the great die-off during the decades after 1492. --Taivo (talk) 01:59, 9 June 2015 (UTC)[reply]

"Proposed Parent Family"

This whole column needs to be removed. It's simply not needed. This chart is, by definition, for top-level language families--those language families that are not demonstrably related to any other. In the currently constituted "Proposed Parent Family" are rejected groupings (Khoisan, Altaic, Caucasian), groupings that are still widely accepted (Niger-Congo), and groupings that have been recently proposed and have not received very much scholarly support beyond the original proposal (Austronesian-Ongan, Dene-Yeneseian). In other words, either we list the "proposed family" as part of the list (based on consensus) or we drop it. Listing fringe or rejected groupings in that column is simply doing a disservice to our readers. There are so few of these anyway that the column is mostly empty. Get rid of the highly speculative or rejected and there is almost nothing left. It's a confusing and somewhat misleading column for our readers. --Taivo (talk) 14:06, 9 June 2015 (UTC)[reply]

I think the information in the column is still (largely) useful, i. e., to provide traditional (and until recently accepted, and still widespread in the literature, and in some cases still accepted by a large number of mainstream scholars) classifications to help recognition and placement, but a better solution would be to convert the column into a group of footnotes, which would also eliminate redundancy and make the code more compact. --Florian Blaschke (talk) 16:17, 7 July 2015 (UTC)[reply]

I agree. This information would be far better in footnotes. --Taivo (talk) 17:15, 7 July 2015 (UTC)[reply]

Location

What does the location column refer to? It seems for some languages it is some ethnographic present (Indo-European: 1492?), but this is not stated. Also, is it consistent across language families?Kdammers (talk) 07:10, 1 January 2017 (UTC)[reply]

@Kdammers: I don't think Indo-European became anywhere as dominant as it is now outside Eurasia before the 18th or 19th centuries, and by most accounts it is still a primarily or at least mostly Eurasia-based language family, with a few languages that can be called exceptions, which does not, however, affect the whole family in general. And apart from Indo-European, I can't even think of any other example where there could really be a debate. That Chinese is spoken outside Asia doesn't mean Sino-Tibetan is not still mainly an Asian language family. So, to be frank, this sounds like nitpicking. --Florian Blaschke (talk) 13:15, 26 May 2017 (UTC)[reply]

I am asking for accuracy and clarity; I don't think that is being nit-picky. At least on the map of North America, the time represented is given, so why can't it be in the chartKdammers (talk) 10:23, 24 June 2018 (UTC)[reply]

External links modified

Hello fellow Wikipedians,

I have just modified one external link on List of language families. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:

Added archive https://web.archive.org/web/20011005193846/http://www.ethnologue.com/web.asp to http://www.ethnologue.com/web.asp

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

This message was posted before February 2018. After February 2018, "External links modified" talk page sections are no longer generated or monitored by InternetArchiveBot. No special action is required regarding these talk page notices, other than regular verification using the archive tool instructions below. Editors have permission to delete these "External links modified" talk page sections if they want to de-clutter talk pages, but see the RfC before doing mass systematic removals. This message is updated dynamically through the template {{source check}} (last update: 5 June 2024).

If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—InternetArchiveBot (Report bug) 23:58, 21 May 2017 (UTC)[reply]

Pictish for unclassified?

Would the extinct Scottish language Pictish qualify for the unclassified list?

This language, although widely considered to be Indo-European (and probably Celtic) is still, at least to some extent, disputed in terms of its exact origins.

It is for the most part considered to be a Celtic language, similar to Welsh. However, some still hold the view that Pictish was non-Indo-European, and a leftover of the Bronze Age languages of Scotland.

JoeyofScotia (talk) 13:53, 28 June 2017 (UTC)[reply]

My mistake. I failed to notice that it was already there! JoeyofScotia (talk) 13:57, 28 June 2017 (UTC)[reply]

"Proposed families" list has zero references. "Mainstream Consensus" column may be original research.

Some references need to be added so that others can check the veracity of these claims. Otherwise there is no way of telling if the entire "mainstream consensus" column is someone's personal opinion. Roidroid (talk) 04:15, 13 April 2018 (UTC)[reply]

Which families do we treat as solidly established and which as proposed?

The list is a mess because there's no canonical list of established families and their components, leading to editors disagreeing with each other and redundancies like proposed families like Penutian being listed in the main list but also (some of) their components, and the same problem with Macro-Jê – although I have now removed the components and inserted Macro-Jê into the main list, there's a new redundancy now in that it is also listed in the proposed list. Pinging TaivoLinguist, Kwamikagami, Maunus and Tropylium specifically. --Florian Blaschke (talk) 21:51, 28 August 2019 (UTC)[reply]

It's not possible to give a canonical list. If we require a reconstructed protolanguage, then there are only a few established families -- not even Austronesian makes the cut, because not all of the Formosan branches are included in reconstructions. Then we get to most people accept but never really demonstrated (Sino-Tibetan?), most accept but there are significant hold-outs, 2ary sources cite advocates but don't have the knowledge to evaluate the proposals, etc. Any list we give is going to be a synthesis of sources. We could give multiple lists from multiple 2ary sources. But most of them won't be reliable, as they won't be experts in the families they classify. Any 2ary list we give is going to be contested by people working in the field who feel it's either too conservative or too provisional. — kwami (talk) 22:06, 28 August 2019 (UTC)[reply]

We do need some sort of consensus classification or there will be constant squabbling and back-and-forth. And the RS and our articles don't help: RS often disagree with each other, and our articles reflect this uncertainty. We just need to make a line between "established" and "proposed" somewhere, or we can only give up and delete this article. --Florian Blaschke (talk) 22:12, 28 August 2019 (UTC)[reply]

We could copy Glottolog. That has the benefit of being conservative, so there shouldn't be many spurious families in the list, though it is a bit loosey-goosey with Papuan families, and it doesn't include some of the families (such as Niger-Congo) that are most frequently mentioned in the literature, though of course those would appear under 'proposed'. Or we could go with separate sources for different continents -- Papua would need to be counted as a 'continent'. — kwami (talk) 22:20, 28 August 2019 (UTC)[reply]

Glottolog is widely considered to be a solid list of language families. I don't think it's quite as "loosey-goosey" as kwami does, but that's just angels on the head of a pin. Anything higher up the food chain than a glottolog family would be "proposed". --Taivo (talk) 04:23, 29 August 2019 (UTC)[reply]

In the general case this does not need to be binary. As long as we have regular sources, we can write an article, but in some cases it may not fall squarely into either the POV of an accepted language family or the POV of a speculative proposal. A few auxiliary article organization features will require a binary treatment, but even there we don't have to use the same cutoff every time, e.g. I'd lean towards more "lumping" with categorization (which is less visible) vs. more "splitting" with navigation boxes (which are more visible). So which sort is this list then more like…? On one hand this is not a landing point for people wanting to know about the affiliation of particular language, on the other it is a quite prominent location for people wanting to have an overview of the state of the field of language classification.

Maybe a good idea might be to try to draw not one firm line but rather two ones: one for "full acceptance" (cannot be "universal", the occasional fringe critics will disagree with just about anything, up to and including the existence of Indo-European), another for "in the works" proposals with some kind of active ongoing support / research (e.g. Austro-Tai, Nilo-Saharan), the third for fringe ideas that only have been even discussed by one or two people. Placement between the first and the second would likely seem less "invalidating" if all clear junk is kept separate in the third section.

Note on sourcing: as far as I know there are very few "secondary sources with lists"; things like Glottolog are tertiary sources. Secondary sources for classification would be things like review articles, or much more often in historical linguistics, more-than-passing mentions in primary sources about something else. E.g. on Macro-Jê, Nikulin (2019) is a recent article that treats the family as accepted and also e.g. lists several MJ etymologies on pp. 122–124 = productively builds on the theory, not just mentions it. --Trɔpʏliʊm • blah 09:35, 29 August 2019 (UTC)[reply]

If we want a "standardized" yet broad-based system (based on more than one reliable source), then a realistic 4-way split would be more linguistically valid: Yes, Probably Yes, Probably No, No. Yes would be a Glottolog entry, since that list is widely considered to be the most conservative (Indo-European, Algic, Na-Dene [Tlingit + Athapaskan-Eyak], etc.). Probably Yes would be above the Glottolog level, but a family that has widespread acceptance (Niger-Congo). Probably No would be a family that has some scholarly acceptance and active scholarship, but its acceptance is clearly in the minority (Nostratic, Nilo-Saharan, Dene-Yeniseian, Austro-Tai). No would be things that are widely rejected and have virtually no work being done on them (Amerind, Austric). --Taivo (talk) 19:34, 29 August 2019 (UTC)[reply]

Numbers of speakers

Where do the numbers of speakers come from? It wouldn't be so bad if they weren't also ridiculously overprecise. I doubt that more than two significant figures for the speakers of a whole language family are ever appropriate – or of interest to the reader, even if they could be verified (by sources – which still has little do with the real world, where even population numbers for well-defined nationalities are probably rarely stable for much longer than a few minutes). Even in a well-defined single language, three significant figures are pushing it. --Florian Blaschke (talk) 18:41, 9 December 2019 (UTC)[reply]

Updated numbers?

Does anyone has access to the latest edition (24) of the paywalled Ethnologue and could update the numbers given here? The List of languages by total number of speakers is up-to-date, so it would be great if whoever had access to the information used in that article could also update this one. --Krissie (talk) 19:58, 14 April 2021 (UTC)[reply]

Eurasiatic and Altaic - "Rejected" or "Controversial" ?

An edit of mine has been reverted, after I amended the "mainstream consensus" of the hypothesised language-grouping Eurasiatic to "controversial" (from "widely rejected") and did so on the groundwork of this being the delineation of the grouping at the applicable article. This was and remains reverted on the basis that Altaic too is thought "rejected" - this language-grouping is likewise under the "controversial" heading at the Altaic article. My proposal is that the "widely rejected" denomination of both groupings in discussion be replaced by "controversial" in reflection of their articles. I am willing to discuss this issue at length and will not go ahead with these proposed alterations without prior approval. JoeyofScotia (talk) 20:01, 26 June 2021 (UTC)[reply]

add isolates to the table of language families

i think it would be good to add also isolates (single language families) to the world language families table with number of speakers. (in the "Language families (non-sign)" section). --Qdinar (talk) 18:36, 26 February 2023 (UTC)[reply]

I disagree. Isolates are not "language families". There should be a separate list for isolates because there are nearly as many isolates as there are language families. If I counted in Glottolog correctly, there are 237 language families and 184 isolates. People coming here to find language families would have to dig through all the isolates and get confused. Create another article on isolates and reference that article clearly here, but don't throw mud in the waters here by including isolates, which are, by definition, not families. --TaivoLinguist (Taivo) (talk) 20:37, 26 February 2023 (UTC)[reply]

i do not think that many people come here to find "language families" meaning only multiple-language language families. "don't throw mud in the waters here by including isolates" - i do not think that i include them, they must be are already included, like by definition, and, indeed, that is so, that is already stated in the very beginning of the article:

The following is a list of language families. It also includes language isolates, unclassified languages and other types.

but really that is not done, they are not really listed in one list. advantage of one list is that [all] items can be sorted by parameters, so single language families can be compared with multiple language families, like by number of speakers.

"isolates, which are, by definition, not families" - i cannot agree with this. i think, isolates are also like form their families, just that families are not explicitly formed and used, because that would be excessive categories. Qdinar (talk) 23:52, 3 March 2023 (UTC)[reply]

Language families and language isolates are not the same. A language family is a group of languages that are related, a language isolate is a language that is not related to other languages. Language isolates are literally by definition not language families. – Treetoes023 (talk) 01:32, 20 August 2023 (UTC)[reply]

no. languages are also typically divided into dialects. even if they are not divided, but consists of one language, they can be considered as a group of one element. and, even a one language is used by many people, and every person uses the same language differently, because languages allow some freedom, and person develop some preferences in using that freedom. also, often even decision of whether to account a language or dialect as a language or as a dialect is an arbitrary decision. yes, they are formally different things by definition, but that distinction is not so important, it does not make "language isolates" radically different things from languages that are not "isolates", and having them only in different lists makes no practical, valuable sense, for purpose of categorising languages. yes, having a capability to see isolates and families separately is an interesting thing, like for purpose of trying to find some other features that unite them, ie, for knowing out which type of languages are typically isolates, and which types are typically not. if you are so fond of the group/single definiton, then, you could create a family for each isolate, and put that family together with other families. btw, the "isolates" name/category was not created for purpose of putting them away from other languages, (that are grouped into families), in the first place, ever. --Qdinar (talk) 22:36, 21 August 2023 (UTC)[reply]

That argument, Qdinar, is not a valid argument. Using that logic you could argue that every speech form can be subdivided into dialects, subdialects, etc. right down to number of idiolects based on number of speakers. Treating language isolates as "families of one" negates the meaning of the word "family" (a group of languages which derive from a common ancestor). This list cannot become a hopelessly meaningless list by including everything. It is a list of families, not languages, and as such should include only families, not isolates. --TaivoLinguist (Taivo) (talk) 01:44, 22 August 2023 (UTC)[reply]

i keep standing at my opinion. " This list cannot become a hopelessly meaningless list by including everything. " - this is a lyish phrase, because i have not suggested to include everything. "It is a list of families, not languages" - i somewhat answered to this. that that intention does not make sense. maybe the first author of the page had such intention. (i have not checked yet, who it is.) but that intention is wrong, in our world. such intention might appear if he/she wanted to just write an article, and he invented this idea based on the term "language family". of course, in space/field of different ideas it is just an idea as other ideas. but, this article should have been "List of language families together with language isolates" instead, that list is what most people are interested in. nobody is interested in "List of language families" because, that list is a lie, it is an incomplete list of language families, because some language families are removed based on unimportant feature of being divided into several languages or not, during history. every language family is also like a language isolate, because it has not strongly enough proven connections with other language families. -- Qdinar (talk) 11:11, 24 August 2023 (UTC)[reply]

As a matter of formal definition, I have to agree with Qdinar that an isolate is a family of one; but it is also true that including them adds to length without proportionate benefit. The proper solution imho is to split off List of language isolates (now a redirect to here) and link it with appropriate annotation. —Tamfang (talk) 21:08, 15 September 2023 (UTC)[reply]

This article is far too much

This article should only include language families, it should not include language isolates and unclassified languages, those should have their own separate articles. I can clear out the entries that are not language families if no one is opposed to the idea. – Treetoes023 (talk) 00:51, 20 August 2023 (UTC)[reply]

I am not opposed. Just basing the current list on Glottolog would make a "family + isolates" list 429 entries long. If you've ever tried to skim that list alphabetically, it is unmanageable. If you sort the list by number of daughters in each entry to make it workable, then you've, in essence, already created two separate and roughly equal lists: those entries with more than one daughter (families) and those with just one daughter (isolates). Unclassifiable languages are a separate list in Glottolog of 121 entries. --TaivoLinguist (Taivo) (talk) 01:40, 22 August 2023 (UTC)[reply]

Language families of Africa

This map is wildly inaccurate, and if it is from 1996, it was widely out of date even then. It shows the west African languages as Bantu! What's going on here? john k (talk) 19:54, 19 January 2024 (UTC)[reply]

Family relation of sign languages

Given the time passed since the addition of this entry in the article (and even earlier) and its citation, could it be worth an update?

Some possible good starting points in order of recency:

Computational phylogenetics reveal histories of sign languages (Abner et al. 2024)

Historical Linguistics of Sign Languages: Progress and Problems (Power 2022)

Historical Linguistics and the Case for Sign Language Families (Reagan 2021) Edittlealittle (talk) 16:07, 14 April 2024 (UTC)[reply]

Creole, pidgins and mixed languages

Should creoles, pidgins and mixed languages be listed under the spoken language families, such as having English-based creoles, French-based creoles, etc. as separate families? — Preceding unsigned comment added by Novioboy (talk • contribs) 03:05, 19 July 2024 (UTC)[reply]

Altaic (again)

I've noticed that several language families on this list have Altaic as their proposed parent family. Altaic is pretty widely rejected, so I've gone ahead and added '(controversial)' next to Altaic where it is listed. The column title says *proposed* parent family, so that's why I haven't just removed those. Let me know if it should be handled some other way. AlexChillOut (talk) 08:09, 24 July 2024 (UTC)[reply]

the memory hole

Of course I'm dying to know why the article's history was censored! —Tamfang (talk) 01:52, 5 November 2024 (UTC)[reply]