Re: [tied] Let dogs have their day too

From: Richard Wordingham
Message: 16088
Date: 2002-10-08

--- In cybalist@..., Piotr Gasiorowski <piotr.gasiorowski@...>
> ----- Original Message -----
> From: Vassil Karloukovski
> To: cybalist@...
> Sent: Monday, October 07, 2002 5:05 PM
> Subject: Re: [tied] Let dogs have their day too
> > at any rate the turkic and iranian words are probably related.
most words for "dog" start with k- (typically kan-) in a very large
number of languges across families.
> The above is an overstatement. I'm aware of a large number of
languages in which the 'dog' word neither begins nor is derivable
from anything beginning with /k/. As far as I know nobody has
actually done any counting, but I doubt if the alleged high frequency
of /k/ is in fact statistically significant. One gets the impression
that "there is something in it" from Ruhlen's listing of "kwan" words
only because the listing is so selective and ignores countless
other 'dog' words in the languages of the world.

I have done a quick survey of the word for 'dog'. I sampled the
Swadesh word lists collected by the Rosetta project ( ) by choosing one word for dog from each
language family recorded other than the artifical languages. Within
each family, I searched the languages in alphabetical order until I
found a word for 'dog'.

I had worried that the rule of one word per family would bias the
results in favour of Amerindian languages. This did not happen –
there were only three American families with Swadesh word-lists!
There is a small bias towards Oceania. There are three minor
problems with the data:
1. Carib and Arawak both seem to have borrowed Spanish 'perro'.
2. Indo-European kwo:n is represented twice – Afrikaans 'hond' and
Haitian 'sje'. As Haitian is, lexically at least, Indo-European,
there is a case for excluding it. ('Afrikaans' precedes it
3. For Ainu the first gloss I encountered was a
reconstruction, 'gita'. I rejected that in favour of the first
attested form, 'seta'.

The results are given below.

Note that V and M are SAMPA-encoded vowels, and that =\ is the
palatoalveolar click.

k: – 7 of the 31. 7/31 = 23%; 7/29 = 24%
kúta , Afar, Afro-Asiatic
kopek, Turkish, Altaic
kO, Bahnar, Austro-Asiatic
kau53, Chinese, Min Nam, Sino-Tibetan
koira, Finnish, Uralic
kaso, Galela, West Papuan
ukhM&, Aceh, Austronesian

h: - 3 of the 31. 3/31 = 10%; 3/29 = 10%
hO , Amto, Amto-Musan
hutt, Kamiyahi, Hokan
hond, Afrikaans, Indo-European

s: - 3 of the 31. 3/31 = 10%; 2/29 = 7%
seta (Proto-Ainu *gita), Ainu-Japan (Shumushu dialect), Language
sje, Haitian Creole French, Creole
su22nak55, Thai, Tai-Kadai

Affricates – 3 of the 31. 3/31 = 10%; 3/29 = 10%
tz'i, Ach?Rabinal, Mayan
txakur, Basque, Basque
psane, Aekyom, Trans-New Guinea

b: – 2 of the 31. 2/31 = 6%; 2/29 = 7%
bóló, Baka-Pigmy, Niger-Congo
bIs(i), Bagirmi, Nilo-Saharan

p: - 2 of the 31 – both from Spanish?. 2/31 = 6%; 1/31 = 3%
péero, Arawak, Arawakian
pe:ru, Carib, Carib

v & w: 2 of the 31. 2/31 = 6%; 2/29 = 7%
wuruwad.e, Anindilyakwa, Australian
vEm, Bauzi, Geelvink Bay

d: 2 of the 31. 2/31 = 6%; 2/29 = 7%
dev, Hmong Daw, Hmong-Mien
duma [ actually, = wolf], Elamite, Isolate

l: 2 of the 31. 2/31 = 6%; 2/29 = 7%
VlOwO:u, Ama, Left May
la, Abkhaz, North Caucasian

n: 2 of the 31. 2/31 = 6%; 2/29 = 7%
na:y, Kannada, Dravidian
nepere, Au, Torricelli

Miscellaneous – 3 of the 31. 3/31 = 10%; 3/31 = 10%
Ought '=\kwe' to be counted with initial /k/?
=\kwe, Kung-ekoka, Khoisan
mitakeu, Lavukaleve, East Papuan
yaury, Alamblak, Sepik-Ramu

The initial consonant is /k/ (or very similar) in 23% of the cases.
If one adds in /h/ and the affricates, for affricates often derive
from /k/, one obtains 13/31, or 12/29 if one discounts /ps/, between
42% and 41%. In fact, of the sample, we know that 'hond', 'sje'
and 'seta' derive from words beginning with gutturals, so words now
or once beginning with gutturals make up at least 9/31 = 29%, or 8/29
= 27% when double-counting of words derived from PIE kwo:n and
Spanish perro is removed.

I would therefore say that there is a large proportion of languages
in which the word for 'dog' begins or derives from a word that began
with /k/. I am not sure how statistically significant these results
are; totting up the statistics of word-initials would be a lot more
work. Does anyone have distributional data to hand? It may need to
be corrected for vowel-initial words, for I have taken the first