False cognates

[Shemtov]

False cognates are words in the same or different language(s) that are similar in form and meaning but have different sources.
Post your favorite ones here:

Welsh and Hebrew have the same feminine singular third person pronoun: [hi].

[Alessio]

In Italian and French the 2sg personal pronoun is "tu". In Hindi it is "tū".

[Click]

In Italian and French the 2sg personal pronoun is "tu". In Hindi it is "tū".

These are real cognates because Hindi is also Indo-European.

[atman]

Here are a few I remember:

"de" and :zho: "de" (both mean "of")
"bad" and :hye: "vad"
"day" and "día"
"hair" and :hye: "her"
"theos" and Nahuatl "teotl" (both mean "god")
"duo" and :zho: "dui" (two)
"kephale" and "kepala" (head)

But the most impressive one is probably:

have and habeo

"Have" is actually cognate with Latin "capio" (I take), while "habeo" is cognate with English "give"!

[nzk13]

Well, relatively recently I thought of
:heb: צד [sˤad] (side)
side

I don't really remember many offhand, but I have noticed plenty of words in Classical Hebrew/Judaeo-Aramaic which are similar to some of some other languages I know, like English. And, every so often, I can wonder if it perhaps got into IE through Greek or Latin borrowings, though I realize that that is pretty unlikely.

[Alessio]

In Italian and French the 2sg personal pronoun is "tu". In Hindi it is "tū".

These are real cognates because Hindi is also Indo-European.

They are far apart from each other, however. Apart from that, there is no other correspondance between these languages.

[nzk13]

Here are a few I remember:

"de" and :zho: "de" (both mean "of")

Also a similar thing found in Aramaic, the prefix "ד" /di/ (which I think comes from the word "די" which means the same, and also seems to be used as the conjunction "that", but "די" seems to be used in later Judaeo-Aramaic mostly as the conjunction, leaving the prefix "ד" for possession.) (the vowel is sometimes reduced to a very short /e/-like sound), which I suppose means the Genitive, as in (right-to-left): "ארעא דכנען" The land דCana`an (of Cana`an). Seeing as Semitic langs are not IE, I suppose this is probably not cognate (though that's a pretty cool coincidence, eh?).

[Click]

In Italian and French the 2sg personal pronoun is "tu". In Hindi it is "tū".

These are real cognates because Hindi is also Indo-European.

They are far apart from each other, however. Apart from that, there is no other correspondance between these languages.

So what? That doesn't show that 'tu' and 'tū' aren't cognates.

[Xing]

The textbook example is perhaps dog in Mbabaram and English. Among short, monosyllabic function words, one would not be surprised to find accidental similarities - like le, masculine definite article in French, and definite/specific article in Samoan. Or ya (Malay/Indonesian) and ja (various germanic languages) for "yes".

[Shrdlu]

[...]Or ya (Malay/Indonesian) and ja (various germanic languages) for "yes".

... and possibly comanche!

[nzk13]

In Italian and French the 2sg personal pronoun is "tu". In Hindi it is "tū".

These are real cognates because Hindi is also Indo-European.

Just to add to this, Hebrew 2SG pronouns are /ʔat'tɔ/ when referring to a male, and /ʔat/ when referring to a female. /ʔat/ (sometimes /ʔant/) is also the Judaeo-Aramaic 2SG, although I think it may be gender-neutral there.

[Shemtov]

In Italian and French the 2sg personal pronoun is "tu". In Hindi it is "tū".

These are real cognates because Hindi is also Indo-European.

Just to add to this, Hebrew 2SG pronouns are /ʔat'tɔ/ when referring to a male, and /ʔat/ when referring to a female. /ʔat/ (sometimes /ʔant/) is also the Judaeo-Aramaic 2SG, although I think it may be gender-neutral there.

Also, in Japanese it's /anata/

[Xonen]

In Italian and French the 2sg personal pronoun is "tu". In Hindi it is "tū".

These are real cognates because Hindi is also Indo-European.

Just to add to this, Hebrew 2SG pronouns are /ʔat'tɔ/ when referring to a male, and /ʔat/ when referring to a female. /ʔat/ (sometimes /ʔant/) is also the Judaeo-Aramaic 2SG, although I think it may be gender-neutral there.

Also, in Japanese it's /anata/

None of these look like obvious cognates of the IE tu, though, or at least I'm not seeing it... The resemblance between /ʔant/ and /anata/ is there, I'll admit, but nonetheless, I wouldn't start reconstructing Proto-Japono-Semitic just yet.

However, first-person pronouns beginning with /m/ and second-person pronouns with /t/ or a similar sound actually does occur in a surprising number of language families; see here, for instance. This has been explained variously as some form of sound symbolism, an indication that the families are actually distantly related, or even as relics from Proto-World. Or, you know, coincidence. There's probably no way to prove any of these hypotheses, but it's a nice little factoid to be aware of, I guess.

[Ithisa]

/anata/ for "you" is not etymologically a pronoun. Modern Japanese has all but lost its pronoun system (/anata/ is originally a noun a.nata meaning "the one over there", Early Modern Japanese even had /konata/ /sonata/ for "me" "you", the word /kanata/ 彼方 still exists) and the original pronouns as far as is attested are wa/ware for the first person and na/nare for the second person. Note that Old Japanese does have true pronouns: wa/ware is declined irregularly (thus differently from nouns) and is not further analyzable.

[Shemtov]

In Italian and French the 2sg personal pronoun is "tu". In Hindi it is "tū".

These are real cognates because Hindi is also Indo-European.

Just to add to this, Hebrew 2SG pronouns are /ʔat'tɔ/ when referring to a male, and /ʔat/ when referring to a female. /ʔat/ (sometimes /ʔant/) is also the Judaeo-Aramaic 2SG, although I think it may be gender-neutral there.

Also, in Japanese it's /anata/

None of these look like obvious cognates of the IE tu, though, or at least I'm not seeing it... The resemblance between /ʔant/ and /anata/ is there, I'll admit, but nonetheless, I wouldn't start reconstructing Proto-Japano-Semitic just yet.

However, first-person pronouns beginning with /m/ and second-person pronouns with /t/ or a similar sound actually does occur in a surprising number of language families; see here, for instance. This has been explained variously as some form of sound symbolism, an indication that the families are actually distantly related, or even as relics from Proto-World. Or, you know, coincidence. There's probably no way to prove any of these hypotheses, but it's a nice little factoid to be aware of, I guess.

In most Amerind the first person pronoun has /n/ in it, which gives some weight to the theory that they are one family.
Interestingly enough, this is also true for the Semitic and other so-called members of the Afro-Asiatic family, Basque, and some Caucasian languages, like Lak.
Again, it's not enough to reconstruct proto-Amerind-Cacauso-Vasconic-Asiatic, but it's interesting.
Note that based on the Medieval work The Travels of Benjamin there was a belief that the people of the Caucasus had a connection to a Semitic-speaking people known as the Gargishites.
However, it does lend some weight to the theory Baque is part of the proposed Dené–Caucasian family.

[Thakowsaizmu]

Here are a few I remember:
"de" and :zho: "de" (both mean "of")
"duo" and :zho: "dui" (two)

The <de> you used here for Chinese isn't really much of a false cognate. Although one of the functions of <de / 的> is on the surface somewhat similar to the Spanish <de>, the pronunciation isn't close, and the grammatical functions are nowhere near close. Aside from sometimes marking possession, the <de / 的> particle is used with adjectives as well, among other functions. Likewise, it is not the only particle that is rendered in Pinyin as <de>, and the other two <de> particles don't even come close to the Spanish.

Also, <dui> is not Chinese for two. <er / 二> is the numeral two, and <liang / 两> is used for counters. I am not sure where you got <dui> as two for Chinese, but your source was a little off.

[Khemehekis]

However, first-person pronouns beginning with /m/ and second-person pronouns with /t/ or a similar sound actually does occur in a surprising number of language families; see here, for instance. This has been explained variously as some form of sound symbolism, an indication that the families are actually distantly related, or even as relics from Proto-World. Or, you know, coincidence. There's probably no way to prove any of these hypotheses, but it's a nice little factoid to be aware of, I guess.

Here's my theory.

I read an article on why "mama" and "tata"/"dada" exist in so many unrelated languages. Mothers hear their babies playing with sounds and think that "mama", the first sounds babies typically play with, refers to them. "Dada"/"tata" comes second, and the parents assume the child is trying to address or refer to his/her father. All around the world, parents have listened to their children babbling, experimenting with the first two consonants they can say, and made these into words.

If /m/ is a typical first consonant and /t/ a typical second consonant, speakers all around the world would think /m/ the most natural sound for a first-person pronoun and /t/ the most natural sound for a second-person pronoun.

What do you think?

[Ithisa]

Here's my theory.

I read an article on why "mama" and "tata"/"dada" exist in so many unrelated languages. Mothers hear their babies playing with sounds and think that "mama", the first sounds babies typically play with, refers to them. "Dada"/"tata" comes second, and the parents assume the child is trying to address or refer to his/her father. All around the world, parents have listened to their children babbling, experimenting with the first two consonants they can say, and made these into words.

If /m/ is a typical first consonant and /t/ a typical second consonant, speakers all around the world would think /m/ the most natural sound for a first-person pronoun and /t/ the most natural sound for a second-person pronoun.

What do you think?

Also interesting is the tendency of Japanese to associate nasal sounds with subjective feelings (compare さみしい /samisii/ and さびしい /sabisii/, both describing "lonely" but the first can only be used for the emotional sense); this may be cross-language. /m/ sounds personal and /t/ sounds distant for some reason.

[atman]

Also, <dui> is not Chinese for two. <er / 二> is the numeral two, and <liang / 两> is used for counters. I am not sure where you got <dui> as two for Chinese, but your source was a little off.

Oh, there's a simple explanation for that: I don't speak a word of Chinese . Plus I cited from memory, making errors doubly likely.

And if "er" is the numeral two, I have another bogus IE etymology: Classical :hye: "erku".

Note to the curious reader: Classical Armenian "erku" is a 100% regular reflex of PIE *dwóh₁. Yes, Armenian regularly changed PIE/Proto Graeco-Armenian *dw into (e)rk, among countless other crazy sound shifts.

However, first-person pronouns beginning with /m/ and second-person pronouns with /t/ or a similar sound actually does occur in a surprising number of language families; see here, for instance. This has been explained variously as some form of sound symbolism, an indication that the families are actually distantly related, or even as relics from Proto-World. Or, you know, coincidence. There's probably no way to prove any of these hypotheses, but it's a nice little factoid to be aware of, I guess.

Here's my theory.

I read an article on why "mama" and "tata"/"dada" exist in so many unrelated languages. Mothers hear their babies playing with sounds and think that "mama", the first sounds babies typically play with, refers to them. "Dada"/"tata" comes second, and the parents assume the child is trying to address or refer to his/her father. All around the world, parents have listened to their children babbling, experimenting with the first two consonants they can say, and made these into words.

If /m/ is a typical first consonant and /t/ a typical second consonant, speakers all around the world would think /m/ the most natural sound for a first-person pronoun and /t/ the most natural sound for a second-person pronoun.

What do you think?

I think your theory isn't bad, but the problem with the m/t pattern is that it's found in most language families in Northern Eurasia, and in few other families in the rest of the world. So we'd want a better, more systematic explanation.

IMO the most likely scenario (bearing in mind that personal pronouns are frequently borrowed across different language families, especially during periods of extensive contact) is that the m/t pattern pronouns were simply diffused across the Eurasian steppe in prehistoric times.

Therefore, it's interesting to note in the WALS map that the only featured North Eurasian language without m/t pronouns is Ket, a Yeniseian language. According to well-received recent research, Yeniseian languages are spoken by peoples who migrated to North America and then came back to Eurasia. Did Ket "fail" to conform to the m/t trend because its speakers were "on vacation" in North America when this sprachbund was active?

[Xonen]

/anata/ for "you" is not etymologically a pronoun.

That doesn't disqualify it from being a false cognate of pronouns in another language, though. After all, the whole point of a false cognate is that it only looks like a cognate if you don't know its actual history.

---

If /m/ is a typical first consonant and /t/ a typical second consonant, speakers all around the world would think /m/ the most natural sound for a first-person pronoun and /t/ the most natural sound for a second-person pronoun.

I think you may be assigning too much value to the terms "first" and "second" here. AFAIU, the reasons for them are purely historical (i.e. what order some ancient grammarians just happened to start listing pronouns in), rather than having anything to do with what order children acquiring language start using pronouns in. Also, I seem to recall reading somewhere that mastering pronouns tends to happen rather late in the language acquisition process, so presumably it shouldn't depend on the order of acquiring phonemes (since that should mostly have happened by then). My knowledge on this subject is a bit sketchy, though, so if someone can find a good source that contradicts me, feel free to present it.

EDIT: Well, I did some (more) googling, and found this:

Unlike many areas of language acquisition, the specific order of pronoun development has been studied extensively (Owens, 1996). Much of the research suggests that pronoun development is variable, although there is some agreement on basic pronouns, such as I, it, and you. (Chiat, 1986). According to Roseberry-McKibben and Hegde, normally developing children begin using first pronouns, such as “I” and “me” at 2-3 years of age, “you,” “they,” “us,” and “them” at 3-4 years of age, most pronouns, including possessives by 4-5 years of age, and all pronouns by 5-6 years of age (Roseberry-McKibben and Hegde, 2000).

So it would appear that children do start using first-person pronouns first, but it's still at an age where they should certainly have mastered more consonants than just /m/.

---

/m/ sounds personal and /t/ sounds distant for some reason.

Yeah, that's essentially the sound symbolism hypothesis. One explanation I've seen is that pronouncing /m/ causes you to purse your lips kind of inward while /t/ includes launching air forwards from the tip of your tongue... But again, whether or not some ancient population of humans would have been on some level aware of this enough to base (what eventually became) their pronouns on it is just not a question we can answer.