Jump to content

Talk:False cognate/Archive 1

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia
Archive 1

Are loanwords considered to be a type of cognate, or not?

This point needs to be more clearly explained right at the top of the article, since some of the example pairs in the "false cognate" list strike me as quite possibly being related by recent borrowing from one language to another. Throbert McGee (talk) 17:56, 17 February 2009 (UTC)

to have

One false cognate that I think is interesting is the verb "habeo" in Latin and "to have" in English and "haben" in German. They have the same meaning, all languages are indo-european, but it is not the same root, the root of have/haben is the same of "capio". I don't know if it fits here, but it's worth saying.Bruno Gripp 01:52, 26 January 2006 (UTC)

That's really worth noting, thanks :-) I was just about to "correct" something... From the article I got a false impression that all the three words are said to have different roots, but now I see the slash there. --213.138.138.89 16:01, 6 June 2006 (UTC) (K Valtakari)
There seems to be some feeling against this fact. I added have vs avoir and haben vs habere to the examples and they were immediately removed. I see no good reason for this, so I've put them back. A strange thing to hush up... garik 01:53, 27 July 2006 (UTC)
On the other hand, I do agree that one can really overdo the examples.
But this particular example happens to be one of the most frequently cited and explained to illustrate the concept (in fact, it is one of the few mentioned in the Wikipedia article Cognate and in the Wiktionary entry False cognate), because Germanic haben/have and Romance haber/avoir mean basically the same and are used in the same grammatical way as auxiliaries to form the perfect aspect (e.g., English to have done is equivalent to Spanish haber hecho). Germanic and Romance languages are both Indoeuropean, so on first sight these words are apparently credible candidates to be cognates; but in fact they are not because we also know that Germanic h is not a reflex of Romance h, and in reality these seemingly related verbs come from different PIE roots (Germanic haben/have comes from PIE *kap-, and true Romance cognates are not haber/avoir but seemingly unrelated words like Spanish captar; while Romance haber/avoir comes from PIE *ghabh-, and true Germanic cognates are not haben/have but seemingly unrelated words like English give). That is, if the article should be left with just one example to clearly illustrate what a false cognate is, I think this example should be that one, because it is the perfect paradigm that has all the ingredients (phylogenetically related languages, similarly-looking words, same basic meaning, same grammatical usage, but... we know for sure they are not cognates and we also know what the true cognates of those words in those languages are and, paradoxically, how seemingly unrelated they actually look like). I can't believe some people have been deleting this prime example from the article, when it should be given a whole prominent paragraph to explain it in detail. By comparison, many other examples in the current article between totally unrelated languages look contrived and even outright dismissible on first sight—I mean, who with a minimum linguistic background (let alone with a minimum knowledge of the involved languages) would even consider that English bullshit and Chinese bù shì (不是 "(it) isn't, (it) is not true", which is not a noun like bullshit and not even one word but a simple sentence involving two of the most basic Chinese words: the negative marker "not" and the copula verb shì "to be", which are taught in lesson 1 of Chinese for Foreigners 101) could conceivably be cognates? Not to mention that words from unrelated languages (such as English and Chinese, or Arabic and Japanese) that are truly related (such as English tea and Chinese chá) are not properly cognates (words that have evolved from the same origin in a common ancestor language), but merely loanwords. Uaxuctum 16:59, 23 January 2007 (UTC)
I agree entirely about habere/haben being a classic example that ought to be included. I wish to point out that my entry bullshit/bù shì was an intentionally extreme example. Someone with no knowledge of one of the languages might hear the word from the other language used and, upon learning its meaning, think there was a connection, whereas someone with even the slightest bit of knowledge of both languages would know instantly that there was no etymological connection and that the similarities were sheer coincidence. I think illustrations such as this are useful in drawing attention to the extremes to which some people, with insufficient understanding of the histories of the languages involved, go in claiming connections between words for which there isn't in reality even a hope of a connection. I've seen people insist recently that habere must be related to haben and English baker must be related to Russian pekar on the grounds that it's obvious. Examples like mine serve to show inarguably (well, except by the deluded) how wrong intuitive notions of "obviousness" can be in establishing etymological connections. —Largo Plazo 17:14, 23 January 2007 (UTC)
I'm not a linguist, but I have read in some places that haben and habere come indeed from the same PIE root *kab-/kap- (latin capere also would come from this very same root). Although being false congnates is perhaps the most accepted theory, I think it should be explained also this possibility (perhaps as it is already mentioned above, there should be one whole section explaining this example).--Xtv - (my talk) - (que dius que què?) 17:32, 20 January 2008 (UTC)
Sources, please. Anyway, Latin habere, unlike capere, simply cannot be derived from the root *kap-: The sound correspondences don't fit. Florian Blaschke (talk) 21:54, 7 June 2010 (UTC)

Many

Article gives English "many" and Korean "manhi" as an example of a false cognate.

I proposes that this may be a true cognate, in the light of Nostratic hypothesis. Let's start from English. English "many" came from Middle English "manye", Old English "manig", reconstructured Common Germanic "managa", Proto Indo-European "menegh", which meant 'many', or 'copious'. Reflexes in other Indo-European languages include Irish "minic" and Welsh "mynych", which mean 'frequent', descended from Common Celtic "minig". Also Russian "mnog(мног)" and Old Church Slavic "munogu", which mean 'abundant'. This is well-established and you can find it in PIE references.

On the Korean side, -i in "manhi" is adverbial suffix, so the root is "manh(많)". In Middle Korean this root is written "manhâ" (â for arae-a, first a is long tone). There are Written Mongolian "mandu" and Chuvash, a Turkic language, "mъnъ", which mean 'big', also attested Old Japanese "mane-si" meaning 'many', but replaced by "opo-si" already in the old time (from Old Japanese "opo" meaning 'big') thus results in modern "ooi(おおい)" but still meaning 'many', and Evenki form "man" meaning 'crowd'. This allows us to reconstruct Proto Altaic root "mana(ha)". This is more controversial, but can be found in literatures: Mongolian-Japanese relation pointed out by Ozawa Shigeo who compiled extensive Mongolian-Japanese dictionary, and also noted by Korean scholar Samuel Martin in "Lexical evidences relating Korean to Japanese", with sound correspondences and all. (Actually, since "m" and "n" is same in all reflexes, there's not much to do.)

This two reconstructed protoform can give rise to tentative Nostratic root "managha", and the rule that Nostratic /a/ gives PIE /e/ and loss of final vowel is regular. (See Allan Bomhard's book "Indo-European and the Nostratic Hypothesis") To get Altaic form you need to assume /gh/ becomes /h/ between vowels, which is reasonable.

You may disagree. But I think there are reasons to believe that m-n- for "many" is likely to have very ancient root, and English "many" and Korean "manhi" may be a true cognate.

Sanxiyn 01:04, 17 Jan 2005 (UTC)

I proposes English "hole", Yucatec Maya "hool" (pronounced same) to be used as an example instead of "many" vs. "manhi", which is hostile to Nostratic hypothesis. Any opinion? Sanxiyn 03:28, 17 Jan 2005 (UTC)

Geez, maybe we could add ...(although some nostraticists think they might be related.)... in that case...

Thanks, Damian. I thought I was the only Westerner in the world to use the similarity between many and manhi as a mnemonic device. I had know idea it was a legitimate linguistic concept, with a name and all! Ed Poor

That's an interesting bit about manhi and many. I noted a long time ago that with Swedish många (many) and Tagalog mga (abbreviation for manga). In Tagalog and other Philippine languages, mga is used to indicate the plural; ang aso (the dog) and ang mga aso (the dogs). --Chris 9 July 2005 05:31 (UTC)
Hmmm, the Swedish word is related to the English, anyway...

Oops! I got tripped over a false friend in the typo (know) in the above sentence. The irony is spinning around in my tummy like an eel.


That's not a false friend; that's just a homophone. A false friend is a homophone that crosses language boundaries. But thanks for inadvertently pointing out the similarity among FCs, FFs, and homophones; I might include those in the next version of the article. --Damian Yerrick

Obrigado gozaimasu

Does anyone know the relationship (false cognate/ borrowing) between the Japanese arigato and the Portuguese obrigado? They both mean "Thanks." I seem to recall that the Portuguese were the first westerners in recorded history to visit Japan. Any ideas? If anyone knows, please add it to the appropriate entry... Steve Rapaport

Last time I checked, I was told that the Japanese borrowed that word from the Portuguese. --Damian Yerrick
If it helps, supposedly 'arigato' is a combination of 'ari'(there is-有り) and 'gato'(difficulty-難う). Meaning there is a strong ethical custom among the Japanese to return favors, like 'I owe you one' thing.--Jondel 09:52, 20 Jun 2004 (UTC)
BTW , Supposedly the Japanese word 'Tabacco' came from Portugese. Perhaps the Japanese 'Baka' and 'Bakero' ( 'stupid')came from 'Vaquero'(?) .--Jondel 00:24, 23 Jun 2004 (UTC)
Is bakero a real japanese word? I couldn't find it at Jim Breen's online dictionary?
You are not thinking of the similar word Bakayarou where -yarou is a pejorative suffic?
http://en.wikipedia.org/wiki/Baka gives no examples of the "Vaca" etymology,
the kanji the word consists of mean "deer" and "horse", but that could be a later spelling.
I don't know if there are examples of the word in japanese, prior to the portuguese-japanese trade routes.
I think this FAQ might be of use. It addresses the obrigado coincidence. No mention of baka, but if it's written with kanji it's probably not foreign, "tabacco" notwithstanding (also addressed there). Lusanaherandraton 15:26, 1 August 2005 (UTC)
Baka ("stupid") is old Japanese (possibly from Sanskrit moha "unenlightened", but this is conjecture), the deer-horse kanji were retrofitted. yarou is a derogatory term for "person", so bakayarou is "stupid person" and bakero is a colloquial contraction. Jpatokal 11:40, 12 October 2005 (UTC)

I concur that "arigato" and "obrigado" are not related. But the similarity occurs to so many people maybe it belongs in the list on this wikipedia entry. I have a question about the Japanese/Arabic "anta". I don't know Arabic, but does "anta" actually exist in Japanese as a relative of the second-person pronoun "anata"? I've never encountered it.

Yes, it does. It's an (colloquial?) abbreviation of "anata", mainly - but not only - used by women. It'a not considered as polite as "anata". "Anata" also means "dear", as in "sweet-heart". The kanji for "anta" and "anata" is the same (貴方), but normally hiragana is used.

El dinnero

Is there any true cognation between English "dinner" and Spanish "dinero"?

No. "Dinner" comes from Latin disjejunare, to break a fast. (French dîner and déjeûner come from the same root, but were formed at different times.) "Dinero" comes from Latin denarius, a Roman monetary unit. -phma 04:41, 11 Mar 2004 (UTC)

that other type of false cognate

This may be a misuse of the term, but the meaning of false cognates I was brought up with were words from different languages that appeared similar, but had very different meanings. The classic example, of course, is embarassed in English and embarasada, or pregnant, in Spanish. Is this a false friend? If not, what do we call this and where does it belong? Italo Svevo

I would say embarrassed and embarazada(with a z) are false friends.
They possibly share the same etymological root,
but since the english word is borrowed from french or latin, they are not true cognates.
Interestingly, false friends in closely related languages are often true cognates.
Examples are English Leek, Swedish Lök (Onion). English Fang, German Fang(catch).
(I think this info should be added to the page.)
No, those are true cognates but false friends. Cognates don't have to have the same meaning, they just have to be related etymologically, as are embarrassed and embarazada. However, it wouldn't be surprising for an English speaker encountering embarazada for the first time to assume it means embarrassed. That's why they're false friends. --Largo Plazo 17:27, 8 January 2007 (UTC)

Gaijin

Is there any possible relation between the Romany/Calo word 'Gadjo' and the Japanese word 'Gaijin', both of which sound very similar and mean the same thing (I'm guessing not -- compare to other 'G' words like hebrew 'Gentile' and American 'Goddamfurner' that also mean the same thing). Only reason I am not sure is that the Romani (gypsies, zigeuner) roam so far and wide they could have theoretically picked up a word from almost anywhere... so I won't just discount it and add it in.


To be added to the list  : Japanese : Anata ( you in English ), Arabic (and Hebrew?) Anta ( also you in English ). --Jondel 09:48, 20 Jun 2004 (UTC).

Gadjo and Gaijin is also similar to the hebrew(or yiddish?) word Goy, meaning non-jewish.
Japanese even has the word Anta, it is a word used by girls and women, derived from Anata.
From a discussion at http://www.zompist.com/chance.htm, concerning sound resemblances between different languages: "Of course it's worse than that, since there's semantic leeway as well. Goyim means 'nations' (from goy 'nation'); though it's metaphorically used for 'non-Jews', it's certainly not an exact semantic match for a word meaning 'foreigner'. (Gaijin is a borrowing from Chinese wairén 'outside-person'.) I can't emphasize enough that inexact matches greatly multiply the chances for finding random resemblances." W=G seems a strange sound shift, but French from Germanic often shows a similar pattern, W=Gu. (Ward, Garder,; Wit, Guide,; (Swe)Vante(Mitten), Gant(Glove).)

Japanese greeting

Another Japanese question - twice recently I have heard a Japanese greeting that sounds very similar to the French "bonjour". Is it just my Western ears, or is there a relation? — Paul G 15:59, 8 Oct 2004 (UTC)

Was this 'Dommo' (Dom-mo)? How was it used?--Jondel 08:57, 3 Jun 2005 (UTC)
Here is what I could think of: ohayou, konnichiwa, kombanwa, ossu, doumo... -Iopq 01:21, 6 November 2005 (UTC)
I hear this all the time and there's no relation at all. If you pronounce "Kon nichi wa" in a lazy way, it comes out like "Kn'shwaa" which really does sound a little like "B'jouaaaa". But while the French phrase means "[Have a] good day", the Japanese means "This day is [a good/bad day, isn't it?]". -- 210.162.163.144 07:08, 28 August 2006 (UTC)

Reverted Edit

"...a false cognate stems from a comparable source but has come to have different meanings in two languages..." This seems to be completely wrong. A false cognate comes from different sources that have, purely by chance, evolved inte words which look very similar both in meaning and appearance. I revert your edit, since I neither could understand fully what you meant, nor believe it is correct. (What do you mean with "comparable source"?) I requote myself: "Interestingly, false friends in closely related languages are often true cognates. Examples are English Leek, Swedish Lök (Onion). English Fang, German Fang(catch). (I think this info should be added to the page.) " Are examples like these, what you are thinking about?

List of False Cognates

The "List of False Cognates" links to a "List of False Friends", =S, maybe we should make another page with a list of false cognates (such as en/jp: sô/so, nay/nai etc...)

Sanskrit

Where does the Sanskrit word come from, is it really unrelated to the common PIE root? http://www.etymonline.com/index.php?search=name&searchmode=none gives Sanskrit nama.

I believe the word was added as unrelated to the Japanese, which is probably true. I removed it, and ask also about the following lines:
  • Hebrew shesh (six) and Persian shesh (six)
    • There is a widespread theory that the Semitic languages (as a whole) are distantly related to Indo-European. If true, this would be an instance of distant cousins falling together, which is a different topic.
  • English house and Hungarian ház (house, block of flats)
    • Is it certain that the Hungarian is not a loan-word from German Haus?

On another topic, there are several distinct English meanings of chop; one of them (in chop suey) is derived from Chinese. Which of these senses does the Uzbek word represent. Septentrionalis 15:08, 28 September 2005 (UTC)

NB "шесть" (Shest') is Russian for "six". I doubt it is related to "shesh". --MacRusgail 18:14, 10 March 2006 (UTC)
The Russian shest' is certainly related to the Farsi (Persian) word shesh, since Russian and Farsi are both Indo-European (and in fact, their counting numbers 1-10 have a few other quite obvious correspondences). Throbert McGee (talk) 14:58, 17 February 2009 (UTC)
Hungarian ház is related to Finnish koti, which is more similar to the reconstructed proto-Finno-Ugric root word, while German Haus comes from proto-Germanic *hūsan. That way the connection looks a lot less plausible, doesn't it? Florian Blaschke (talk) 21:49, 7 June 2010 (UTC)
Although Germanich *hūsan is of unknown origin, inital *h- often corresponds to Non-Germanic initial *k-. Not that I believe these words are related, anyway, but it isn't such a good exaple of contrary evidence. 惑乱 Wakuran (talk) 16:14, 9 June 2010 (UTC)

False Cognates Within a Language?

I've added trawl and troll, two English words with extremely similar (though not identical) meanings, but with completely different etymologies. If they were in different languages there would be no question that they would be false cognates. Is there any reasonable argument for why they should not be so considered, just because they're in the same language?

Zsero 2 October 2005

I don't know the answer to that, but pen and pencil are similarly unrelated etymologically. Just thought I'd share. – Quadell (talk) 13:50, 2 October 2005 (UTC)
Of course you can have false cognates within a language, just like you can have real cognates within a language. --BluePlatypus 03:55, 20 November 2005 (UTC)
True cognates within a language being called doublets, right? --Damian Yerrick 17:34, 26 December 2005 (UTC)

The word troll might be used like trawl sometimes, but I believe such usage is in error, arising from the similar pronunciation of the words. See also then instead of than and make due instead of make do. 41.185.88.83 (talk) 23:52, 22 January 2011 (UTC)

True cognates

English name and Japanese namae are true cognates (from Nostratic *lemne)) --Nixer 15:54, 12 November 2005 (UTC)

if you accept Nostratic, that is. Not exactly a mainstream proposition, but sure, not impossible either. dab () 16:00, 12 November 2005 (UTC)
Japanese form from proto-Altaic *lihomona, English form - from PIE *(e)nomen. By the way, English know is an even more deep cognate with Chaneese kan to inquire, investigate, to see, look. I think we should place here the true false cognates only.--Nixer 16:13, 12 November 2005 (UTC)

Also Korean manhi and English many are true cognates (from Nostratic *manga). Though in Korean it may be borrowing from Japanese manki. By the way, many in Russian is mnogo--Nixer 20:09, 12 November 2005 (UTC)

  • Arabic akh (brother) and Mongolian akh (brother) ma be true cognates. Mongolian one is from proto-Altaic aka (elder brother), which is from proto-Eurasiatic haka - elder brother, which is from Borean haka - elder relative. In proto-Afroasiatic thas root is in form hk(k) - elder relative.--Nixer 13:02, 15 November 2005 (UTC)

Source

http://www.linguistlist.org/issues/12/12-1906.html I'm reverting -Iopq 23:30, 11 January 2006 (UTC)

Arigato

"Japanese arigato (thanks) and Portuguese obrigado (thank you)"

Actually I have heard that the Japanese is a loanword - any comments on this? Domo is one of the native terms. --MacRusgail 18:11, 10 March 2006 (UTC)

No, it's not a loanword. See gairaigo#One gairaigo misconception. --Kusunose 03:05, 11 March 2006 (UTC)

additions

Added: English persecution (from PIE root sequ - to follow) vs Russian presechenie (from PIE root sec - to cut). --Nixer 04:48, 6 May 2006 (UTC)


False False Cognates

I think English Cold and Russian Holod are both derived from Proto-IE: *g'elǝ- (compare Latin gelu, gelidus), so they would therefore not be false cognates. Twalls

Look here:[1]. This root does not have Slavic derivative.--Nixer 19:16, 17 July 2006 (UTC)
Look here: [2] it says it's most likely from that root and discusses other theories -Iopq 13:03, 21 October 2006 (UTC)


Some of the false false cognates are evidenced by contributors inserting without research or references. The following have no references:

  1. English "name" and Malay "nama". Indeed they are cognates because many Malay words are of Dutch, Spanish and Portuguese origin due to colonisation of the Malay Peninsula by these European "colonial powers". Furthermore, many Malay words are of Sanskrit and Arabic origin. Therefore, Malay-Latin-IndoEuropean cognate words are highly frequent.
  2. The other is the English "canteen" and Chinese "canting". There is evidence that "canting" in Chinese did not exist as a term until exposure to English. I am sure, like many other languages, many European words have become cognates in Chinese due to the widespread and international use of English.

Unless someone produces references and evidence for these two as false cognates, I propose that these be deleted from the list.

I am wary that contributors are willy~nilly inserting false cognates without any reference. The list needs a clean-up, because evidently many in the list have no reference or etymological scholarship. If someone supports me, please insert the clean-up banner for this article.

Hence Jewish Anderstein (talk) 19:16, 3 May 2009 (UTC)

False friends

False cognate is sometimes also loosely used to mean false friend. Is this even worth stating? I mean, it's not just loose, it's downright incorrect. garik 01:55, 27 July 2006 (UTC)

Arabic

Because they arrived in the English language about the same time, the words for Sheriff (from English "Shire Reeve" an officer of considerable power in the 11th century and the Arabic "Sharif," denoting a ruler. The latter was known to Crusaders shortly thereafter. The English word is sometimes listed as "derived" from Arabic, which is untrue (false cognate).

Very poor article

This article is based on wild speculation and overzealous dismissal of very real possibilities. For instance, to say off hand that the numbers SIX, SEVEN in English (Indo-European_ are accidentally, coincidentally similar to the corresponding Hebrew (Semitic)numbers SHESH, SHEV'A is perhaps going too far. Trying to invent so-called 'roots' of an imaginary language called "Proto-Indo-European" and "Proto-Semitic" using interpolations and permutations to attempt to prove this point is unscientific and obtuse. It would be more prudent to entertain the possibility that certain so-called false cognates are not so false after all.

It's entirely scientific to the extent that the reconstructions are based on establishing consistent patterns that account for the changes claimed from the ancestor language to the descendant languages, which are normally based on consistent patterns in the relationships between words that obviously correspond in attested languages. There isn't anything unscientific about noticing that Latin initial {f, p, c} + l (flamma, plenum, clave) (flame, full, key) correspond to Italian {f, p, ch} + i (fiamma, pieno, chiave), Spanish ll (llama, lleno, llave), and Portuguese ch (chama, cheio, chave), inferring that language ordinarily evolves in regular ways, and deducing patterns for languages earlier than the ones for which we have examples. What isn't scientific is failing to account for the lessening degree of confidence we have in existing reconstructions the further back we go to try to create new ones from them, and the degree to which the concepts of semantic and phonological relatedness are stretched by many of the people claiming to have demonstrated vastly prehistoric relationships. --Largo Plazo 17:46, 8 January 2007 (UTC)

Some of these "false cognates"

Can anyone tell me how the English evaporate and the Russian ispar'at' can be false cognates? They seem to be about a million miles from each other. I'm also a bit cautious about delete and udalit. Can anyone verify (from reliable sources) that these two are really false cognates? --さくら 08:58, 20 October 2006 (UTC)

Although I don't have a good Russian etymological dictionary available, my guess just from looking at the word ispar'at' (испарять) is that it's actually a LOANWORD from "evaporate" via calqueing -- I mean, some Russian words just scream calque and this is one of them. That is, the Russian prefix is- (ис-) translates the Latin prefix e(x)-, and par (пар) means "steam" -- thus, both words literally mean something like "to steam out." The only question would be whether this Russian word was calqued directly from scientific Latin evaporare, or if it came by way of a modern Western European language such as French. At any rate, the entire entry needs a much clearer definition of the distinction (if any) between cognates and loanwords. Throbert McGee (talk) 15:42, 17 February 2009 (UTC)
All two words need to be false cognates is to not be cognates, but to look enough like each other (and occupy enough common semantic space) that people might think they are. The first condition is not so bad - we have a pretty good idea that 'evaporate' and 'isparat' are not cognates; but the second is a matter of taste, really. There's no objective way of verifying it. But the two you've removed are a bit dubious to my taste too. garik 09:35, 20 October 2006 (UTC)
delete and udalit' are not cognates. udalit has root dal "far", so it is literally "to remove", "move away", it is used in computers "udalit' fayl" "delete a file".--Nixer 14:48, 21 October 2006 (UTC)
So they are "false cognates" then, because they sound and look somewhat similar but have different roots. They are not cognates but a false ones. Is that right? Dieter Simon 22:18, 21 October 2006 (UTC)
But the thing is that they don't sound or look the same! --さくら 11:08, 22 October 2006 (UTC)
In fact the English word sounds like "delit".--Nixer 12:09, 22 October 2006 (UTC)
IMO delete sounds like "de-leet" and udalit "oo-duh-lit". You're right, there is a small similarity. But only a small one. --さくら 13:03, 22 October 2006 (UTC)
What's the difference between leet and lit?--Nixer 13:08, 22 October 2006 (UTC)
"Leet" is (supposed to depict) a long sound, "lit" short. --さくら 16:31, 22 October 2006 (UTC)

You said there is a small similarity between "udalit" and "delete". Any less similar than these four examples taken from the actual list in the article?:

  • Bangla kaata, English cut
  • Coptic per, Etruscan pera
  • Egyptian maRar, Japanese miru, Spanish mirar
  • English cheek, Russian scheka

It depends what you mean by "similar". Dieter Simon 20:29, 23 October 2006 (UTC)

Well I'm not a linguist, but all of those examples you listed do sound similar to me. One should know how to pronounce both words in order to judge if they truly are false cognates. --さくら 12:50, 24 October 2006 (UTC)
True enough, although occasionally spelling is, perversely, a better indicator of relatedness than pronunciation, since it can reflect an earlier pronunciation that's more obviously related (the spelling of knight, for example, is a better clue of its connection with German Knecht than its pronunciation [nait]); the issue is complicated further by the fact that many real cognates look very little like each other (e.g. English cow and Latin bos). Human families are a useful analogy: my brother looks very much like some members of our family, but doesn't look much like me. On the other hand, he also looks quite a lot like some people who have no connection to my family. We might call these people 'false relatives' of my brother. To judge whether two words 'truly are false cognates' is as subjective as judging whether two unrelated people 'truly are false relatives'. I might think unrelated person A looks just like my brother; you might disagree. There's no definitive answer. Saying someone is a 'false relative' of my brother would be no more interesting than saying 'doesn't that guy over there look a bit like Prince Philip?'; and to say something is a 'false cognate' is merely equivalent to saying: 'You think it's a cognate? Well it's not.' garik 13:48, 24 October 2006 (UTC)
Hear, hear. It is a matter of what you see as a similarity. Dieter Simon 22:30, 24 October 2006 (UTC)

Day

After seeing that someone had removed it, someone else had readded it, and someone then re-removed it, I'm restoring "day". One of the removers wrote "see definition ... both have roots in PIE". "False cognate" doesn't imply "not having roots in a common language", it means "not having a common root in a common language". The American Heritage Dictionary gives the PIE root "agh-" for "day" and the root "deiw-" for "diary".

Agree. When this was first added, I said to myself "That's gotta be a mistake." But I looked it up, and it is a great example. The OED says: "DAY: ... in no way related to L. dies...". It should stay. -- Dominus 15:35, 9 November 2006 (UTC)
Interesting to which root of the above related Russian den'?--Nixer 16:23, 9 November 2006 (UTC)
Because of the palatalized initial consonant I'm going to guess "diary" (cf. Latin "diēs") but I could be entirely wrong.

Removal of Latin "opti", Russian "ochi"

Someone added Latin opti and Russian ochi ("eyes") to the list of examples. Since the two words don't look alike, no one would guess them to be cognates. Therefore, I removed them. (Opti isn't even a Latin word in the first place. Opti- is a prefix taken from Greek optos, ops.)

I suspect the person who added them got the story confused, because ochi does look like Italian occhi = "eyes", from Latin oculi. But in that case, it seems likely that the Russian and Latin terms are true cognates, with unremarkable transformations going back to the reconstructed Proto-Indo-European. Likewise, English "eye" comes from Old English éage, cognate with Old High German ouga. If someone wants to reintroduce Russian ochi, that person should explain here why it isn't related to its Indo-European look-alikes. --Largo Plazo 17:19, 7 January 2007 (UTC)

R. Ochi (sing. oko) is a true cognate to Lat. oculus. -iopq 16:50, 10 January 2007 (UTC)

Are you sure?

Are you sure that Latin Deus and Greek theos are not real cognates? As the languages are indeed related, these words seem so to me too. And German haben and Latin habere? I just say they seem very cognate, but maybe I´m being deceived and they are indeed a very good false cognate. However, Finnish sinä and Turkish sen? In the [Uralic languges] article, the word "you" is said to be a word orginating from the very same Uralic proto-language. I just want to make sure you know what you´re talking about. Besides, I see no point in this many examples, especially when some of them seem doubtful. Do they really make the article better?Laurelindë 19:41, 6 February 2007 (UTC)

Yes, we're pretty sure. This is more or less the point of the concept of false cognates: some things look very much like cognates, but aren't. Deus is related to Greek Zeus; haben is related (I think) to capere—compare pairs like hund/canis where Germanic h represents Latin c. I do agree, however, that the examples could be pared down rather. garik 20:31, 6 February 2007 (UTC)
On checking my dictionary just now, I was surprised to find that Greek theos and Zeus are not cognates! Zeus and its Latin cognate deus both come from an IE root meaning "to gleam," while theos is from an entirely different IE root meaning "to storm." Thus, Gk. theos is a false cognate with Gk. Zeus and Lat. deus. (I think there's enough conceptual similarity between theos and Zeus to add them as a false pair.) Throbert McGee (talk) 15:16, 17 February 2009 (UTC)


"They seem very cognate"—isn't that the point of the article? :-) Yes, haben is related to capere in the same way as Hund to canis, as Garik points out, and also as hundert/centum (hundred) and haupt/caput (head). Habere is related to geben (to give) in the same way as hortus to Garten (garden) and hospes to Gast (guest). —Largo Plazo 21:59, 6 February 2007 (UTC)
Turkish isn't an Uralic language. The Finnish pronoun "sinä" comes from earlier *tinä. Muhaha 17:30, 19 February 2007 (UTC)
Deus and theos should be real. The whole indo-european hypotheses relied on words like deva(ind)/deo(lat) .24.203.68.10 (talk) 14:56, 26 January 2010 (UTC)

Stop with the examples already.

Well, that's my opinion. I don't see how the article is helped by more examples. The point is to explain false cognates. I don't think they add enough value to the article to warrant the amount of discussion going on here regarding various words. If someone really wants to have a list though, I suggest it be moved to a seperate article (List of false cognates or similar). --BluePlatypus 03:58, 20 November 2005 (UTC)

Great idea. Do it. -Iopq 08:21, 28 December 2005 (UTC)
I couldn't agree more. A lot of articles suffer from this problem (onomatopoeia, for example). A separate list of cognates is not really needed, as far as I am concerned: Wikipedia is not a general knowledgebase. — mark 19:42, 28 December 2005 (UTC)
I've moved this section to the bottom since it's still pretty relevant. I've also added the following suggestion to the top of the examples section:
I'm not convinced that a separate article with a list would be of much benefit to anyone (although a small part of me quite likes the idea); I think we're probably better off just pruning the list in this article, or at the very least not adding to it. What do people think? garik 10:53, 1 March 2007 (UTC)
<!-- Before you add a fresh example, please ask yourself if it's really necessary.  Wikipedia
 is not about creating endless lists.  Rather than adding an example, you could almost certainly serve
 this article better by making improvements to the main body.-->
The problem really is that this is another one of those open-ended lists which is going to go on forever if we let it. As someone comes across a pair of similar-looking or -sounding words with similar meanings in different languages they shove them in. Can you imagine how many there must be in all the world's languages? And really I agree,a separate list doesn't alter the fact, the best thing is just give us a flavour of the thing so that we get an idea of what "false cognates" are like. Yes, a pruned list of the most apposite and appropriate examples is best. But even that is probably going to be a problem, who is to sort it? Dieter Simon 02:32, 2 March 2007 (UTC)
A short list should consist of classic examples where the non-relationship is uncontroversial and concisely demonstrable, and the concise demonstration should be provided. —Largo Plazo 14:29, 2 March 2007 (UTC)
A great example! Someone just added English much and Spanish mucho. That the "ch" is coincidence is apparent as soon as one sees that mucho goes back to Latin multus. That these words only look like cognates is undisputable and easy to demonstrate. —Largo Plazo 00:30, 3 March 2007 (UTC)
And I should have mentioned that "much" developed the affricate /tʃ/ late, from /k/, being a shortening of "mickle", ME "mycel". —Largo Plazo 20:12, 25 September 2007 (UTC)

English"strange" and Russian "странно"

Don't they both come from Latin extraneus, meaning "out of place"? Devil Master 14:31, 4 June 2007 (UTC)

"Strange" certainly does come from extraneous, via French. My Russian etymological dictionary is curiously silent on "странно", but I would assume that it's derived from страна, which has a different root from extraneus (which is based on extra). garik 14:55, 4 June 2007 (UTC)
According to the Dal' dictionary (толковый словарь Даля) странный does indeed originate from страна. Dnquark 06:21, 20 September 2007 (UTC)
Yes, strannyi ("weird") is from strana ("a country"). Furthermore, Vasmer's Etymological Database says that strana -- now meaning "a country" in the sense of "nation-state," but at one time having the more specific implication of "FOREIGN country" -- derives from storona, meaning "side" or "party." (Cf. modern Russian storonnik, "partisan.") This original sense of foreignness or other-side-ness inherent in strana helps explain why the adjective strannyi has the meaning "strange; weird; odd." Throbert McGee (talk) 17:18, 17 February 2009 (UTC)
I suppose this is cognate to either English "stretch" or "string" in its original meaning "something that stretches", compare "prostranstvo" - space, "prostor" - wide space, room, "prostirat's'a" - to cover wide area.--Dojarca (talk) 10:00, 20 September 2009 (UTC)

Why this article is useless

The user "Zsero" asked me why I deleted the words I did. Answer: because this page is a mockery to linguistics and the world would be better off without it. for example, you idiots are questioning the credibility of every dictionary in current use (Webster's, Oxford...) cognates like The English word "day" the Latin word "dies" and Spanish word "día"? what is wrong with you glory seeking VAGRANTS? THIS is from the WIKIPEDIA article on "day" "The term comes from the Old English dæg, with similar terms common in all other Indo-European languages, such as dies in Latin and dive in Sanskrit." You people have no idea what you're doing, this is the kind of thing that makes me regret Wikipedia being open for ANYONE to edit. This article is riddled with ridiculous claims about KNOWN COGNATES as well as words nobody would be ignorant enough to even think in the first place that they were cognates. You're all just throwing things in here when you say yourself that "you're not sure" as the person below commented about "to have" maybe, just maybe if you're not sure about something you shouldn't put it on this page for people to read and be MISINFORMED, the point of Wikipedia is to educate not to randomly post you're opinions as if it were a "Blog". THIS ISN'T MYSPACE PEOPLE.--Christianus 02:23, 21 September 2007 (UTC)

The OED says, of "day":
In no way related to L. dies.
So perhaps you should consult the dictionary from time to time yourself. This error on your part, and your ranting, severely undermine your credibility. -- Dominus 02:48, 21 September 2007 (UTC)
On the other hand, you are right about neat/nitidus, which I just removed. Perhaps it would be more constructive to provide detailed citations below, as I just did for "neat". -- Dominus 02:55, 21 September 2007 (UTC)
Given the contradiction between this article and the one on "day", how did you wind up concluding that the "day" article is a fault-free, scholarly piece, while this article is the one with an error? The person who wrote that in the "day" article is the one who has made the mistake. (I see now that someone deleted that incorrect assertion a couple of days ago.) And you should be really careful about citing references you haven't first bothered to look at. The Oxford English Dictionary says about "day", "In no way related to L. dies". Merriam Webster Online unabridged edition shows completely unrelated etymologies. —Largo Plazo 13:15, 23 September 2007 (UTC)

Eng. "neat" / L. "nitidus"

I removed these because the OED claims they are cognate. Under NEAT (a.) it has:

[< Anglo-Norman neet, neit, variants of net (see NET a.).]

and under NET (a.) it has:

[< Anglo-Norman net, neth and Middle French net clean, morally pure (12th cent. in Old French in senses 1 and 2, a1483 in sense 3; French net) < classical Latin nitidus NITID adj. ...]

Someone should probably go over all the items and add citations. -- Dominus 02:54, 21 September 2007 (UTC)

Thank you for bringing that error to my attention, albiet easily made. Even still this page is in need of attention, namely in the cases of "hut" and "it".--74.242.90.177 03:58, 21 September 2007 (UTC)

Ger. Kreuz / Rus. крест (krest)?

Are you certain about these words? "Kreuz" is, anyway, ultimately derived from Latin "crux", just like English "cross", etc. and it seems probable Russian/Slavic words like "krest" would be, as well? 惑乱 分からん * \)/ (\ (< \) (2 /) /)/ * 22:50, 27 September 2007 (UTC)

Russian "krest" derived from "Christ".--Dojarca (talk) 17:22, 1 January 2008 (UTC)
Okay. 惑乱 分からん * \)/ (\ (< \) (2 /) /)/ * (talk) 15:44, 20 February 2008 (UTC)
Krest came from Greek Christos ("the anointed one") and originally entered the Slavic languages with the meaning "Christ." By metaphoric extension, the name Krest later came to mean "the Crucifixion (of Jesus)" and from there the sense development went from "the Cross" to "a crucifix" and finally "any cross-like shape". Throbert McGee (talk) 16:18, 17 February 2009 (UTC)
In modern Russian, by the way, "Christ" is Xristos (Христос), not Krest (Крест). However, the word for "to christen" is krestit' , yet the newly baptized person becomes a xristyanin ("Christian") -- not to be confused with krestyanin, "peasant"! Throbert McGee (talk) 16:36, 17 February 2009 (UTC)

Japanese pan

From the article:

"Japanese pan (bread) and Spanish pan (bread)"

According to the article Japanese words of Portuguese origin pan comes from Portuguese pão. Is it certain that Portuguese pão and Spanish pan are not cognate? I'm removing this example for now. Ossi (talk) 22:54, 22 December 2007 (UTC)

Japanese pan is derived from Portuguese pão, which clearly is a cognate to Spanish pan. That is not a "false cognate". 惑乱 分からん * \)/ (\ (< \) (2 /) /)/ * (talk) 15:45, 20 February 2008 (UTC)
exactly, why is this word here? the word in japanese is written in katakana (パン) meaning it has to be borrowed from a foreign language, and other users have reported that it comes from the portuguese, which is clearly related to the spanish "pan". They have the same origin and meaning, therefore they are true cognates User:Okashii

Etruscan & Sumerian

Etruscan ac (to make, act) and Sumerian ak (to make,act)

Etruscan an (he/she/it) and Sumerian ane (he/she/it)

Etruscan ipa (who, which) and Sumerian aba (who)

Etruscan mi (I/me) and Sumerian ma (I/me) and Korean na (I)


I was under the impression that both languages are considered isolates, so how is it possible to say that some pairing is a true or false cognate? Virginia-American (talk) 03:01, 20 June 2008 (UTC)

Ural Altaic

Finnish sinä (singular you) and Turkish sen (singular you)


These are considered evidence for the Ural-Altaic, Nostratic, and Eurasiatic hypotheses. I don't think they should simply be labeled as "false".

Virginia-American (talk) 13:46, 20 June 2008 (UTC)


Exactly. This whole list has this attitude; just because something has not been proved as a true cognate, it is labelled as a false one. The two are certainly not the same thing. —Preceding unsigned comment added by Okashii (talkcontribs) 00:27, 6 October 2008 (UTC)
As pointed out above, Finnish s comes from an older t. That's the reason why the comparison in misleading. Florian Blaschke (talk) 21:44, 7 June 2010 (UTC)

Similarities between English Name, and Japanese Namae?

Are they false friends, false cognates, gairaigo/wasei-eigo?Moocowsrule (talk) 03:13, 8 September 2008 (UTC)Moocowsrule

They are true cognates. Please look through this talk page. --Dojarca (talk) 09:07, 9 September 2008 (UTC)
The above discussion states that they are true cognates in the conjectured family of Nostratic languages. But common Nostratic descent is far from consensus among historical linguists. --Damian Yerrick (talk | stalk) 14:02, 9 September 2008 (UTC)
Well whether Nostratic theory correct or not, these words no doubt connected at some higher level just as 'many' and 'manhi'. Youy can thace both roots at this database [3].--Dojarca (talk) 15:10, 9 September 2008 (UTC)
Not necessarily. In fact, I'd say the balance of probability falls on the side of their not being cognates. Although I very much sympathise with efforts to reconstruct languages ancestral to the commonly accepted protolanguages, I have yet to see good evidence for the nostratic reconstructions or for any other such superfamily. I'm happy not to include "name" and "namae" on this page — it has long suffered from a superfluity of examples in any case — but the evidence for their being true cognates is slim. garik (talk) 15:29, 9 September 2008 (UTC)
Consider this link: [4] Also note that even in Indo-European hettite name is 'laman' - compare it with proto-Altaic 'lihomona'. The reconstructed Nostratic form is close to 'lemne'. Also consider Proto-Uralic 'nime' (comp. Finnish and Estonian 'nimi', Mordovian 'lem', Nenets and Udmurt 'nim' etc) and Chukchee-Kamchatkan 'nenbe'. Also you can see that this root is traced even further into hypotetic Borean with analogies in Afroasiatic ('nam', 'lam'), Sino-Caucasian and Austric ('lan').--Dojarca (talk) 20:34, 9 September 2008 (UTC)
But this kind of coincidence is depressingly common. Consider Latin habeo and German haben. They're most definitely not cognates. Now, I'm not saying that Indo-European and all the other protolanguages supposed to belong to Nostratic can't have a common ancestor; I'm saying that the superficially convincing examples always trotted out aren't all that impressive as evidence. Mass comparison is a poor method, and coincidences abound. Lyle Campbell has a good paper on it somewhere; it's also worth reading McMahon and McMahon's Language Classification by Numbers. garik (talk) 12:12, 10 September 2008 (UTC)
I quote from Mass lexical comparison:
Lyle Campbell, a linguist specializing in the languages of the Americas and author of a review of Greenberg's book, was able to  
establish a correspondence between the proposed Amerind language and Finnish, and others were able to do so with Latin and many 
languages obviously not related to those of the Americas.
I would also direct you to this page. garik (talk) 14:32, 10 September 2008 (UTC)
Starostin's school, to which the cited site belongs explicitely reject statistical methods, rather aiming on reconstruction. Also relation between Altaic and Indo-European is now rarely disputed. There are too many reconstructed roots for Nostratic.--Dojarca (talk) 21:01, 10 September 2008 (UTC)
Well, I'm sorry, but I think you're missing the point. The point is that a fairly simple application of statistics shows that mass comparison is an unreliable method, likely to produce a high number of false positives. Human intuitions are highly misleading when it comes to spotting patterns, and we need formal statistical methods to make sense of the patterns we seem to see, and to tell us if they're likely to be more than just coincidences. Just because we're highly adept at seeing faces in the clouds doesn't mean they're there. And just because a lot of languages have words that look a bit like "name", if you squint, doesn't mean that these languages are related. garik (talk) 10:32, 11 September 2008 (UTC)
Again, Starostin's school reject mass comparison as a method. They use regular correspondencies between languages.--Dojarca (talk) 13:08, 11 September 2008 (UTC)

Oh, and I'm sorry, but the relationship between the Altaic languages and the Indo-European languages is still highly disputed. For goodness sake: the existence of an Alatic family at all is still disputed. garik (talk) 10:37, 11 September 2008 (UTC)

Further thoughts on the uselessness of this article

The use of term "unrelated" and "false cognate" refers to a lack of documented or culturally/logically traceable etymological descent. All words, of course, were originally formed *without* any cultural reference by humans in overtly various but in many ways similar (in that they were human with human needs) circumstances.

There is little doubt *in my mind* that phonemes such as the "SS" (soft-S) sound are associated with objects in nature which also make that sound (snakes, snow and silk sheets?). So when someone finds a cross-language connection and "scholars" (those slippery sages) deem that the words "false cognate" are applicable, what they are missing is the deeper (perhaps Jungian?) connection in the *origin* (and in the natural reinforcement and evolution... kings and wars be damned) of language itself.

Indeed, I feel that this deeper connection is the only connection worth studying and the one which lack of study seems operative. The exhausting robot-like documentation of language and linguistic descent by traditional etymology often insults the beauty and grace and nature of mankind's quest to be understood. Poets, philosophers and cognitive psychologists have been trying to make this case for years. But the system we have for sanctifying linguists (and other 'ists) seems to bar access to creative thinking.

On knowledge in general: the most important things we know cannot be justified with easily reducible logic. This is not to say that they are leaps of faith or irrational conclusions. Just that they are conclusions arrived at using the available evidence in a manner which is sufficiently complex enough so as to not allow for trivial inquiry. (Ask anyone in love, whether with nature, woman or god, to explain why and there are not volumes on earth to contain this revelation). Fortunately, there are methods of inquiry with which these conclusions can be tested, known as "experimentation" ... thus without necessitating an understanding of "why" a thing is correct, we can infer it's correctness in, at least, a broad range of circumstances. The modern replacement of the experimental model with proof and derivation, whether it applies to linguistics or particle physics, is unnecessarily delaying the the acquisition of knowledge critical to the understanding of our world and ourselves.

Simultaneous (talk) 15:37, 25 September 2008 (UTC)

I'm sorry, but as far as I can tell, this comment is irrelevant to improving the article. Please read WP:Talk garik (talk) 15:45, 25 September 2008 (UTC)
I disagree in part. Is there a reliable source that suggests that onomatopoeia and other forms of sound symbolism might be responsible for some of these false cognates? --Damian Yerrick (talk | stalk) 19:28, 25 September 2008 (UTC)
There will be several. It's well acknowledged among linguists that, for example, the fact words for mother tend to begin with an /m/ in numerous languages is not the result of their being cognates, but of the kind of noises babies make when they're reaching for the teat (reduplicated CV syllables are, for similar reasons, very common in terms for family members: dad(a) and papa are English examples). And again, onomatopoeia is very well ackowledged as a source of similarities. I'll have a look and see if I can get hold of a source. It's probably mentioned in most introductory textbooks on historical linguistics. But this isn't really the point, as far as I can see, of the above comment, which seems to be treating this fact as a major problem with the article. garik (talk) 10:55, 28 September 2008 (UTC)
It is one thing to recognize or identify onomatopoeic words such as "Mama", "Dada", and most likely quite a number of other universal baby-engendered sounds, but the more obscure sounds which are claimed to be results of onomatopoeic developments really need to be sourced. We cannot just point to the similarity of sounds in words and say: "Here it is, here is the evidence". We cannot produce our own research results, but as long we can cite reliable sources which say, yes, this Japanese word has wandered via some of the Chinese or Mongolian languages or dialects, and has finally come to rest in other languages then we must include it and thus show what we have suspected all along. But do not just assume such a development, we do need sources cited - the first principle that should guide us. Dieter Simon (talk) 23:54, 17 November 2008 (UTC)
Anyone *can* point to the similarity of sounds as evidence, but some people have a cultural *aversion* to this. There was a recently published article in Science or Nature of a mathematical analysis that showed how the length of a word inversely correlates to its level of usage, and that this is preserved across all languages, cultures and times that we know of. This is a trivial example, I know. (I still think "s" sounds come from snakes... how could it be proven, I can't even think of an analysis that would satisfy linguists.) —Preceding unsigned comment added by Simultaneous (talkcontribs) 16:17, 4 June 2009 (UTC)

Reorganization proposal

Why not put the examples into categories, something like:

Known to be false: (because both etymologies are known) Gmc. haben, Lat. habeo Engl. day, Lat. dies

Considered false because it's an isolated occurence: Engl. tit, Swahili titi Engl dog, (? one of the Australian languages) dog

Widely-suspected of being onomatopoietc: mama, aqua

Considered evidence for an Indo-European Semitic relationship (i.e. Nostratic): six, seven, a few others

Considered evidence for Ural-Altaic, Indo-Uralic, etc (i.e. Eurasiatic): Finnish / Turkish pronouns, widespread first-person m, second-person t, etc. Indo-European eghom, Chukchee -egom, -egot

Completely unknown status (because both languages are isolates) Etruscan / Sumerian examples

Considered evidence for Proto-World (very widespread distribution, prob. not onomat.) Famous examples from Ruhlen and the Russians. e.g., tik (finger), putV (female organs)


I'm no expert, but this seems to reflect the true state of knowledge/ignorance

Virginia-American (talk) 19:17, 17 November 2008 (UTC)

Once again, as I said in the previous section, first and foremost, you must cite reliable sources. It is one thing to claim this is the true state of knowledge ( someone else may well disprove it as ignorance), but as long as you have the right sources to substantiate what you are saying you are home and dry and it is much more likely the Wikipedian community will accept it .Dieter Simon (talk) 00:20, 18 November 2008 (UTC)
I'd say it's a good idea to group them in that way. Also, it should be pointed out that in some cases the jury is still out (such as day/dies, which _might_ be related etymologically) and in some cases the words might have influenced each other (such as habere/have or isle/island). Actually, what you call an "isolated occurence" is the only _real_ false cognate, and even *gudaga (which turned into the Australian (Mbabaram) "dog") and English "dog" _might_ be related via something like Proto-World (if it existed) *kunu or *kundu... After all, nobody knows where "dog" came from.
www.etymonline.is a nice source. Look up "dog": "...the origin remains one of the great mysteries of English etymology." *gudaga might be a "hound-dog". EXTREMELY improbable, but still possible.
So, a good definition would be "False cognates are pairs of words in the same or different languages that are similar in form and meaning but TRADITIONALLY CONSIDERED BY LINGUISTS TO have different roots."
By the way, it's funny that Campbell compared Finnish to an "Amerind" language. Greenberg, considering Amerind languages to be related to his "Eurasiatic" family, probably smiled at this.
Ruhlen considers "aqua" to be a Proto-World item, as far as I remember.--Physiognome (talk) 16:40, 20 November 2008 (UTC)
I've often wondered if "aqua" is onomatopoiec (?sp), the sound reminds me of gagging, or gurgling or gargling. Same comment for the proposed worldwide etymology that includes "milk" meaning "swallow", throat", etc. Same comment for "snout", snot", "nose", etc. I'm sure I'm not the only person to have thought of this, but I don't know of any publications that mention it, so it stays out of Wiki under the "original research" rule.
What are the dates when Europeans first encountered Aboriginal languages and when "dog" entered English?
Dieter Simon makes a good point. I think a constructive approach would be to show *why* certain pairs are considered false. EG, the American Heritage dictionary of IE roots lists different sets of English words cognate to "deus" and "theos" respectively.

Virginia-American (talk) 17:02, 20 November 2008 (UTC)

The Old English hapax legomenon "docga" ("hound breed") is apparently from 1050, early European sightings of Australia is 1606, but it doesn't appear to be more thorough exploration of the area until much later. All examples of Aboriginal loanwords in Etymonline are from the 18th or 19th century. 惑乱 Wakuran (talk) 22:14, 8 January 2009 (UTC)

Just as an idea - why not list prior forms/prior roots for each of the false cognates? That will at least (1) better illustrate the idea and (2) help stop the gratuitous addition of true or possible true cognates to the list. 72.83.180.99 (talk) 20:47, 23 March 2009 (UTC)

I like the idea of being explicit about why each pair is cognate. All Proto-World arguments like "dog" are fairly irrelevant, since at that time-depth any two words "might" be cognate; we are only interested in the words coincidentally converging to similar form from dissimilar forms.
Speaking of citations, some for the definition of "false cognate" would be good to see too. --Trɔpʏliʊmblah 10:41, 28 April 2009 (UTC)
Physiognome, day and dies cannot even be related, because Germanic (= English) d and Latin d at the beginning of a word is no regular correspondence. Germanic d and Latin f, and Germanic t and Latin d is.
The reason why Campbell chose Finnish (and others Latin etc.) is probably that its history is well-known and many apparent cognates can be shown to be false. But if you consider this kind of outlandish comparisons to validate the method instead of casting doubt on or even disproving it (by way of it seemingly "proving" even more distant relationships, including "Proto-World") you have missed the point, which is to show why Greenberg's methodology is unsound. Historical change of sound, morphology and meanings as well as the strong possibility of borrowing (even over arbitrarily large distances) make such comparisons worthless because the element of chance is too strong (the "noise"). Compare Klingon and Esperanto with "Proto-Amerind" if you will; I guarantee you it is possible to show that Klingon and Esperanto are Amerind languages with Greenberg's method (including vague resemblances and misrepresented or spurious examples). See pseudoscientific language comparison. Florian Blaschke (talk) 21:42, 7 June 2010 (UTC)

Removing the list

How about we just get rid of that big list at the bottom of the page? There's a link to a nice online list of false cognates anyway. We could have maybe three examples, between different pairs of languages, one of which should include English. Everyone game? garik (talk) 11:23, 13 May 2009 (UTC)

I say preserve it on a more suitable Wikimedia project, if anything. Wiktionary's collection of attested etymologies would help solve the verifiability problem. --Damian Yerrick (talk | stalk) 12:03, 13 May 2009 (UTC)
No we should keep the list, it is more comprehensive then the external link anyway, and so is wikipedia material. Le Anh-Huy (talk) 07:37, 17 May 2009 (UTC)
But just being more comprehensive than an external link doesn't make something wikipedia material. The list here just seems to swamp the whole article. I wouldn't dislike it so much if it had an article to itself. At any rate, we could at least remove any that are unsourced (currently almost all of them), and only put them back if we can find a reputable source describing them as a false cognate (or, I suppose, a source simply showing that they're not cognates for cases where the words are very similar). garik (talk) 10:35, 17 May 2009 (UTC)
I like the list because it gives people something to wrangle about. It keeps editor interest. I come back here from time to time to review and research specifically because of that list. Wikipedia is suffering because the constant "castration" of contentious content causes editors to become bored with the articles (and, to some extent, readers to be bored with the content). I forget where I read an interview with Jimmy Wales that basically amounted to just that. —Preceding unsigned comment added by Simultaneous (talkcontribs) 20:29, 18 October 2009 (UTC)
I agree that it should be preserved. But I don't think Wikipedia is the best place for it among Wikimedia projects. It appears to fit Wiktionary's goals better. --Damian Yerrick (talk | stalk) 14:59, 19 October 2009 (UTC)

wahine?

I have left the Hawaiian part as I do not know how it is pronounced, But in Maori it sounds nothing like the Latin in any way shape or form. I found it quite amuzing to see it there.210.185.17.217 (talk) 08:47, 17 January 2010 (UTC)

How is it pronounced in Maori? In Latin it was pronounced /wɑginɑ/ (rather as if spelt "wageena" in English). This is pretty close to how the Hawaiian word is pronounced: /wahine/ (rather as if spelt "waheeneh" in English). Bear in mind that we shouldn't necessarily expect real cognates to sound very much like each other; English hound (/haʊnd/), for example, is actually cognate with Latin canis (/kɑnis/). garik (talk) 12:15, 17 January 2010 (UTC)
The impression I get from the Wikipedia article on the Māori language is that the Maori word sounds quite a lot like the Hawaiian one. That's really not very far from the Latin at all! That said, I'm all for reducing the pointlessly long list in this article. garik (talk) 12:19, 17 January 2010 (UTC)
Indeed real cognates don't always sound alike, However False ones should, otherwise... Where's the percieved connection? Maori, it can be either "wah-he-neh" or "wai-he-neh" depending on where you're from, I think I might of heard one more pronounciation. Perhaps, but they are not near enough that I would have conceived a possible connection. I did however, again forget the fact my particular Latin dialect is slightly different. Taking that into consideration I'm still a bit... iffy, about this "cognate" my friend here agrees so, I guess I shall leave it upto others. And I agree this list is far too long for examples, it should be shortend, I say leave only the ones that sound very, alike.210.185.17.217 (talk) 12:59, 17 January 2010 (UTC)
Actually, I'd say that "wah-he-neh", which I interpret as /wahinɛ/, sounds very alike to /wagina/ (something like wag-ee-nah), which is the classical Latin pronunciation. The two words differ in only two sounds: /g/ vs /h/ and the final vowel. Now, a shift from /g/ to /h/ is a very common change, and be a typical example of lenition (a very similar change occurred in the development of Spanish, for example). And a shift of this sort in a final vowel is not in the slightest bit unlikely either. In other words, based on form alone, these two words look very very much like possible cognates. We have two reasons to suppose that they are not: first, the meaning is not as close as it appears. The Latin word originally meant "sheath". When you consider the distance between the two languages, then the connection becomes even less likely. So they're almost certainly not related; the phonetic form, however, is really very close. garik (talk) 13:44, 17 January 2010 (UTC)


cwmwl / cumulus ?

Anyone can shed some light here ? (Cumulus is a latin name used to describe some puffy clouds scientifically.) 24.203.68.10 (talk) 15:03, 26 January 2010 (UTC)

Hmmm, "cwmwl" means "cloud", apparently. Cumulus is Latin for heap, but when I google it seems that many people think it's a Latin borrowing into Welsh. 惑乱 Wakuran (talk) 14:10, 13 May 2010 (UTC)

baba (rus, jp, Romanian)

Babă (articulated "baba" = "the hag") also means "old hag" in Romanian, and it could be a cognate of the Russian баба since many slavic words leaked into Romanian too. 24.203.68.10 (talk) 15:03, 26 January 2010 (UTC)

Also, I added yiddish "Bubbe" —Preceding unsigned comment added by 72.230.132.87 (talk) 03:03, 28 October 2010 (UTC)

Bengali "Kaata" & English "Cut"

This could be a true cognate, not a false one. The Bengali word is derived from the Sanskrit term "Karta", while the English word is derived from the North Germanic term "kut". Both are Indo-European languages, which could have retained the old term, albeit in modified forms. Kaata is not unique to Bengali either, but also found in many Indo-Aryan languages, including Hindi/Urdu --Maurice45 (talk) 03:03, 27 March 2010 (UTC)

Cut is of unknown origin. Might be Scandinavian. Not enough evidence apart from conjecture. 惑乱 Wakuran (talk) 14:04, 13 May 2010 (UTC)
Besides, Indo-Iranian initial *k doesn't seem to correspond to Germanic inital *k, anyway. II *k seems to correspond to G *h, while G *k seems to correspond to II *g. Cf. Indo-European sound laws. 惑乱 Wakuran (talk) 21:18, 10 February 2011 (UTC)

Hebrew shesh (שׁשׁ) (six) and Persian shesh (six) and Spanish seis (six)

Hebrew shesh and Persian shesh are false cognates, as is Hebrew shesh and Spanish seis, but that does not make Persian shesh and Spanish seis false cognates! Both being Indo-European, they certainly have a common root. --Funper (talk) 22:08, 28 November 2010 (UTC)

All seem to come from *swéḱs, but the references are missing for wikt:שש and wikt:شش. JackPotte (talk) 00:46, 30 November 2010 (UTC)

Hebrew "Meyaledet", Greek "Maia"

Both mean midwife, but it's extremely unlikely that there is any connection. TFighterPilot (talk) 22:26, 15 May 2011 (UTC)

Examples

The Examples section is listed alphabetically by language. Re-listing it by word would be far more user-friendly. 92.25.214.67 (talk) 20:10, 9 July 2011 (UTC)

Written false cognates

What about words that are written similarly and have the same meaning but have different origins? i.e Chinese character for seven "七" looks like an upside down Arabic 7 with a strike — Preceding unsigned comment added by Sumirp (talkcontribs) 10:52, 27 July 2011 (UTC)

Archive 1