LINGUIST List 6.572

Sat 15 Apr 1995

Disc: Comparative method

Editor for this issue: <>

Directory

1. Jacques Guy, Sakao (to do with the comparative method, somewhat)
2. Alexis Manaster Ramer, Binary vs. N-ary Again

In 1974 I did a quick comparative study of the Hog Harbour and Port-Olry
dialects of Sakao (Espiritu Santo, Vanuatu).

Sakao is characterized by having innovated stress, then developed a
large vowel system (11 or 12), lost many unstressed vowels or most of
the contrasts of those that remained. A few examples will make this
clearer than a long set of diachronic rules.

Key: IPA, with:

A = back rounded a, B = (beta), D = <delta>, E = <epsilon>, G = <gamma>,
I = high, central vowel, unmarked for rounding, N = (ng), O = open o,
R = trilled r, o" = (oe), Y = <o-slash>, W = bilabial co-articulation

HH = Hog Harbour PO = Port-Olry Tol = Tolomako Tsu = Tsureviu
PV = Proto-Vanuatu (my tentative reconstruction)

 PO HH Tol PV

ulyG nulyG na Buluku na Buluku my hair
ulym nulym na Bulumu na Bulumu thy hair
ulo"n nulYn na Buluna na Buluna his/her/its hair
ulyn nulyn na Bulu na... na Buluni the hair of the
nYl nYl na Bulu... na Bulun hair of

You get the gist of it. Very briefly, final vowels were lost and the
tonic (usually penultimate) became:

 Posttonic Examples:
 C(aeo) C(iu)
a a E matana -) mDan matamu -> mDEm
 its eye thy eye
e A o" kerena -) GArAn keremu -> Go"ro"m
 its behind thy behind
o O o" roNo -) rOG moti -> mo"t
 hear broken
i o" (1) y lima -) lo"n tsitsi -> sys
 five flay
u o" (1) y luna -) lo"n lumu -> lym
 in it in thee

(1) o" in PO, but Y in HH, thus HH lYn "five" or "in it"

One Hog Harbour morpheme, however, gives a few worries, thus:

PO HH Tsu PV

ho"G hAG seku seku my
ho"m ho"m semu semu thy
hAn hAn sena sena his/her/its
ho"n hAn se na seni na... of the...

Here the PV reconstructions are made on the evidence of Sakao and
Tsureviu (Tolomako has pila- instead of se-). However, HH /hAG/
reconstructs to *sek(aeo) and /hAn/ to *sen(aeo). Some speakers even
have /hOG/ for /hAG/, and that reconstructs to *sok(aeo). So what do we
have there? Vowel assimilation comes to mind: *sena -) *sene
(assimilation), *seku -) *seko (lowering) then, in some varieties, *seko
-) *soko (assimilation again). Natural, it is precisely such frequent
morphemes that "wear out" phonologically.

Yes... but:

PO HH

walDyG walDiG my child
walDym walDym thy child
walDo"n walDYn his/her/its child
walDyn walDin the child of the
alD nwalD child of

(Proto-Worlders rejoice: the Sakao word for "child" is _obviously_
Arabic walad!)

HH /walDiG/ and /walDin/ are diachronically almost impossible: stressed
Sakao /i/ can only arise from *i(aeo). One could account for
these forms, very messily:

PO (- *GWalati- "child"
HH (- ditto, except for forms 1 and 4 from *Gwalati(aeo)-

But why this exception?

The solution is much, much simpler.

Hog Harbour has had the misfortune of having been reduced to writing by
Presbyterian missionaries and some parts of the New Testament have been
published in it. "Misfortune" because

1. they never quite got the vowel system right
2. they used this dreadful system whereby, to avoid resorting to
 diacritics, you make do with... italics! (I also saw an old hymnal in
 which, instead of resorting to italics, two dots were written *under*
 the vowel).

Here is the Hog Harbour spelling system used in the New Testament for
the vowels:

Spelling Phoneme(s)
 a /a/
 _a_ /A/ or /o"/
 o /O/ or /o/
 _o_ /A/ or /o"/
 e /E/ or /e/
 _e_ /o"/ or /Y/
 i /e/ or /i/, and serves for /j/ as well
 _i_ /y/, or /Y/
 u /u/
 _u_ /y/, or /Y/, rarely /o"/

Thus /walDyG/ was spelt "walth_i_c", /ho"G/ "h_o_c" or "h_a_c". Often,
of course, the vowels were not italicized when they should have been, or
were when they should not.

Having thus added confusion, it remained to add injury: the Sakao of the
Hog Harbour New Testament is a mixture of Pidgin Sakao and baby talk.
The Hog Harbour dialect is characterized by the loss of almost all
atonic vowels. Thus:

 PO HH

o"hy nhhy "rifle" *na susuvi (- *suvi "to blow"
o"BDyG nBDyg "my head" *na kWatuku
o"rImyG nRmyG "my noise, thought" *na nrumuku
AwAlk- nlk- "hand, branch of" *na GWele k(i) "tail, extremity of"

The spelling used in the New Testament, however, is sprinkled with epenthetic
vowels, e.g. /nRmYn/ "his thought" is spelt variously "nu-rumun", "nuru-mun",
"nu-rurumun", etc. You seldom hear epenthetic vowels in normal speech, only
in baby talk used by women to small children, and in "formal", i.e. church,
speech. The language of the Hog Harbour New Testament is also pidginized:
it shows none of the holophrastic verbal constructions typical of both
dialects of Sakao. Pehov (/pehOB/), who was secular chief of Hog Harbour
at the time, told me how the missionaries of his childhood used
the realis paucal for all tenses, numbers and persons, even for the
imperative, e.g. /i tmjan/ "you go" indifferently for /i nmjan/ "you
go/went", /i nGjan/ "go! (polite)", /i jan/ "go!".

Finally, to add insult to injury and confusion, children were taught
this mongrelized language as "proper talk" at Sunday school (as I have
it from Pehov).

And that is why Hog Harbour people now say /hAG/ or /hOG/, /walDIG/ and
/walDin/, and will probably continue doing so as long as their language
survives.

Some time ago, while we were discussing comparative linguistics,
the question was raised (by Bill Poser, I think) of whether anybody
ever claims in print that binary comparison is preferable to n-ary
comparison. While I cited one reference at that time, I thought it
might be still be of general interest to adduce one more, which has
just come to my attention.

In his attack on the theory that Japanese is Altaic (and on Altaic
as a whole), Janhunen 1992 argues that the odds of finding apparent
matches simply by chance when Japanese is compared to the four
Altaic languages/subgroups, viz., Turkic, Mongolic, Tungusic, and
Korean, are four times as high as are the odds of finding such
spurious matches when Japanese is compared to just one language,
specifically Korean (which is singled out by Janhunen because there
has been a fair amount of work on Japanese-Korean comparison which
ignored Altaic).

In other words, Janhunen assumes that a 5-ary comparison is four
times as likely to produce matches purely by chance (what I call
'false positives') as is a binary comparison. This, needless to
say, is a fallacy, but there you have it.


Janhunen, Juha. 1992. Das Japanische in vergleichender Sicht.
 Journal de la Soci t Finno-ougrienne 84.145-161.