Appendix talk:Dene-Yeniseian Swadesh lists

Similarities

I don't see many similarities between words in these three languages. 71.66.97.228 07:24, 22 August 2010 (UTC)

Not many will be visible. The feature that really ties Navajo to Ket and Tlingit is the classifiers, such as -ł-. If we compare Navajo to languages such as ÔÓÌÍKQL;',MNA./l;kcm s; and Apache, you would see a lot of similarities. At least you would if the common linguistic examples are used, which, unfortunately, are often specific to Athabaskan languages and not included in a Swadesh list (such as the postpositions such as shił). —Stephen 12:12, 22 August 2010 (UTC)

Thank you; that is very interesting. R. Carlos Nakai told me that he took his father to meet some Athabascans in Alaska and his father said their language sounded like the way Navajo people used to speak years ago. 71.66.97.228 08:06, 23 August 2010 (UTC)

There are still some visible cognates, such as the ones for "stone" (Ket təˀs, Tlingit té, Navajo tsé), "foot," and "mouth." These languages had diverged many thousands of years ago, so it's only natural for them to go through a lot of change. In comparison, languages such as Hittite and Albanian do not share many cognates with most other Indo-European languages, but you can see they are still IE languages if you look at sound correspondences and grammatical structures. Here are some interesting articles you might want to take a look at - http://www.pnas.org/content/95/23/13994.full.pdf, http://www.uaf.edu/anlc/docs/dy_vajda_perspective.pdf, http://www.uaf.edu/anlc/docs/vajda-2008.pdf. Stevey7788 01:44, 27 August 2010 (UTC)

The evidence is still *extremely* weak, even in Vajda's paper, even if it has the support of some old school Alaskan specialists focusing on particular languages who must not have seen sufficiently general comparative arguments elsewhere. No decent reconstruction to speak of. A few extremely strained, unconvincing and random cognates, some very basic typology found in many families (in fact, even then, both closer to surrounding families than to each other), and a supposedly similar classifier system that is also very strained. We've all cobbled together far better coincidences as an exercise in being conservative in our claims in introductory linguistics courses. Never mind horizontal transfer/areal features that don't imply genetic relationship. At the VERY least, can we not just assume this is a language family? 2604:2000:DF08:E800:A1FE:E7DA:E974:AE24 03:11, 26 December 2017 (UTC)

It's clear that you do not have an understanding of how words are linked with others through reconstructed proto-languages after an extremely long period of time. Athabaskan and Yeniseian have been separated by such a great distance and amount of time that it is unlikely that any words will show the resemblance that a tyro like you think they should. That's not how it's done, as you would know if you had a formal education in the matter. So your folksy requests will not be considered. —Stephen ^(Talk) 10:52, 26 December 2017 (UTC)

Came across this again. I do have a 'formal education' in comparative linguistics to postgraduate level, though with specialisation in another region, and I am speaking about *assuming* such a relationship, as in the prior comments. Before I wrote this I had already gone through Vajda's paper and some commentary on it, and it was far more tenuous than I expected. What I wrote was a high level summary, not a folksy complaint, thanks very much. You are making the assumption even more strongly than Vajda does, in fact. If you're assuming that we can assume this is a language family on here then it's clear *you* do not have any such understanding yourself. I am aware, of course, that there are many linguists try to draw such connections across deep macro-families, and by all means it's good work to try and see what can be found. But you're *assuming* this based on extremely weak evidence - far weaker on its face than several language families that have been debunked when far more numerous and convincing similarities - is far more 'folksy'. This is by no means considered established by linguists in general, as you surely know. Assuming this link is valid when the vast majority of the linguists do not is not 'how it's done', it's how it's not done, and lacks of any basic understanding of the misconceptions about language change and probability that have plagued the whole field. Your argument seems to be 'because we can only possibly find very weak evidence, weak evidence is what is required'. I assume that in order to reconstruct Proto-World and *assume it is correct* we need basically zero evidence then. If we only have weak evidence like this, then we never jump to such conclusions. That doesn't even reach a basic level of critical thinking. By all means consider it interesting, and worth researching further. And I don't know if you read my comment as aggressive in some way, but don't you insult me like that out of the blue. Harsimaja (talk) 19:00, 27 May 2020 (UTC)

KET/NAVAJO COGNATES

• ENGLISH: Ket, Navajo

• LARGE: qaˀ, tsoh (g/q => j/z)
• SNAKE: tìɣ, tłʼiish
• LEAF: əə, atʼąąʼ
• GRASS: daan, tłʼoh
• MEAT: īs, (a)tsįʼ
• BLOOD: sūl, dił
• MOUTH: qō, (a)zééʼ (g/q => j/z)
• EAR: ogde, (a)jaaʼ (g/q => j/z)
• FOOT: kiˀs, (a)keeʼ
• HAND: laŋat, (á)laʼ
• NECK: kə̄qt, (a)kʼos
• STONE: təˀs, tsé
• DAY: iˀ, jį́
• MOON: qīp, ooljééʼ (g/q => j/z)
• STAR: qòʁ, sǫʼ
• LAKE: deˀ, tooh
• SKY: ēs, yá
• SALT: təˀ, áshįįh

— Stevey7788 20:00, 17 September 2010 (UTC)

Which based on Vajda's paper form the majority of the argument given, with a bit of typology and some classifier affixes. How can you look at half of these and not see a poor coincidence, or the other half and not see how random and unalike they are? A computer could come up with a better correspondence between any two random languages. And the only 'systematic' sound changes have to be of the form [sound 1 OR sound 2] -> [sound 3 OR sound 4], which increases the probability of a hit massively. More than half of these have a consonant in common at best, and even then with this laxity. How is this not on par with some of the most unconvincing macrofamilies out there? 2604:2000:DF08:E800:A1FE:E7DA:E974:AE24 03:16, 26 December 2017 (UTC)

How about a relationship among these:

Ancient Greek: φώγω (phṓgō)

Armenian: բոկ (bok)

Germanic: *bakaną

Phrygian: βεκος

Slavic: *bažiti

Ukrainian: багати (bahaty)

Do you see the relationship? Do they look like coincidences and random selections? These are the final results of a lot of work. You only see the results, not the work that produced them. Reconstruction is a science, not a game of bingo. You expect to see relationships similar to German komen and English come between languages that diverged eight or ten thousand years ago or longer. The results should not be close at all; they should look like poor coincidences and random selections. —Stephen ^(Talk) 12:17, 4 September 2018 (UTC)

No, but that's not the point. There is a *mountain* of other evidence of consistent sound changes across hundreds of cognates between any two, morphology, etc. We're talking about the full amount of evidence, not random bits of it. I don't get your logic: this is like the old joke of saying there must be invisible man next to you because you can't see him, and "You wouldn't see him if there were, would you?" You need a *positive*, strong, argument for this. The arguments tying Ket and Tlingit etc. are based on a list of proposed cognates about as long (a few dozen) and mostly very tenuous as we can do for fun between any two languages due to sheer coincidence, on a classifier with a similar phoneme, and the rest seems even weaker. In fact, in order to increase the 'hit' rate, the mathematically convenient trick of {A, B} > {C, D}, which drastically increases the probability, is still overall inconsisten. Yes, that *could* be all that might be left after well over 10,000 years, but that's still not evidence. Not to mention that after requiring a demonstration that the association is beyond coincidence, it is necessary to argue for actual genetic relationship. Altaic has *FAR* more convincing parallels, as do e.g. Vietnamese and Chinese, but we know these to be due to borrowing or areal features. There are many intelligent and well-educated researchers very attached to their macrofamily hypotheses, but ultimately this should rest on convincing argument. The pile of evidence for this is orders of magnitude away from that for IE, and far too close to what we'd expect by chance, but it simply cannot reasonably be assumed to be from a former areal connection, let alone family, even if Vajda is a great teacher and a great linguist, and it would be super cool if this were the case. Harsimaja (talk) 19:10, 27 May 2020 (UTC)

stem noun table

Stephen asked me to have a look -- I throw this in here; not sure how you guys set up lists here. The whole deal works best for the so-called "stem nouns", e.g. the few true monosyllabic nouns. Seb az86556 00:20, 23 October 2010 (UTC)

Navajo	Sarcee	Chipewyan	Beaver	Carrier	Sekani	Ahtna	(English)
chaaʼ	tsxa	tsáá	chááʼ	tsaa	tsáʼ	tsaʼ	beaver
tsé	tsáh	tΘee	chée	tΘee	tsee	tsʼees	stone
tłʼoh	(guu)tłʼowí	tłʼog	tłʼógi	tłʼooh	tłʼóó	tłʼogh	grass
łid	tłi	łʊr	łʊd	łʊd	łʊd	łet	smoke
tó	tútii	tuu	chúu	tuuh	chúu	tuu	water
kʼos	kʼos	kʼoΘ	kʼos	kʼvs	kʼos	kʼos	cloud
sǫʼ	sʊn	tΘʊ́n	sʊ́n	sʊm	sʊm	sonʼ	star
kǫʼ	kuu	kon	kʊ́n	kʊn	kon	konʼ	fire

Additional terms?

Can additional terms of relevance to the American Southwest be added to this list, such as "maize," "turquoise," or "pottery"? 71.66.97.228 18:31, 7 November 2010 (UTC)

Since this is a Swadesh list, no, they can't. You could, nevertheless, create a list of your own that would focus on this type of cultural vocabulary. It might be interesting to see too. --Petusek (talk) 17:56, 24 September 2015 (UTC)

Ket dialects

It should be made clear which Ket dialects are being listed or, more precisely, which words come from which dialects. The Ket column looks rather messy in its current form.--Petusek (talk) 17:48, 24 September 2015 (UTC)

The Ket words should be written in Ket Cyrillic alphabet. If we get enough Ket Cyrillic terms, then it will be worthwhile to write a transliteration module for Ket (Lua program) which will transliterate it automatically and consistently. Somebody just needs to find the time to do this. The transliteration will be fairly easy, but first we need Ket words written in Ket Cyrillic. —Stephen ^(Talk) 04:46, 25 September 2015 (UTC)

Transcription, transliteration

It is not clear which of the Ket forms are phonetic transcription, which are merely transliteration/romanization. These things should be sorted out. Also, you might want to consult this page. --Petusek (talk) 17:53, 24 September 2015 (UTC)