LINGUIST List 5.1269

Fri 11 Nov 1994

Misc: linguist, Chi-square analysis of body/one, Doc. programs

Editor for this issue: <>

Directory

1. , RE: 5.1226 "linguist"
2. Celso Alvarez Caccamo, Terminology
3. , -body/-one
4. vcook, Re: 5.1229 Sum: Doctoral programs in Applied Linguistics

Message 1: RE: 5.1226 "linguist"

Regarding Tony Hall's amusing suggestions, if the benighted masses ask us now
what we do as 'linguists', imagine the questions aroused by our proclaiming
ourselves 'tonguesters'!
 --Jules Levin
"Slavic linguists are the most amusing linguists around" --Roman Jakobson

Anyway, linguistic nomenclature always
depends on the Linguistics Nomenklatur.

Celso Alvarez-Caccamo
lxalvarzudc.es

Lou Burnard posted some data on the -one/-body taken from a sample from the
British National Corpus. Just for fun, I ran a chi-square analysis on the
data. Chi-square is a nice and simple analysis method that tells us if a
matrix is non-randomly put together or not. So, here are the data, again:

 written spoken total
 anyone:anybody 57:31 17:32 74:63
 someone:somebody 72:38 36:64 108:101
 no-one:nobody 18:57 2:18 20:75
 everyone:everybody 81:43 12:23 93:66

and if we reformulate them:

 ANY- wr sp total

 -one 57 17 74
 -body 31 32 63

 total 88 49 137

and run a chi-square test, all the tables, except "NO-", test significantly
(0.001) positive for non-random distribution. (The "NO-" table has too
few occurrences in one of its cells for chi-square to be reliable anyway.)

ANY- X2 = 11.4643
SOME- X2 = 18.1925
NO- X2 = 1.86200
EVERY- X2 = 10.8296

So, chi-square tells us that there is some (ir)regularity in the data.
However, chi-square gives no explanation. There is something there; now try
find out what it is!

(It is surprisingly difficult to determine this sort of frequency
distribution thing by impressionistic inspection -- one of the three
matrices below could represent a random distribution. The other two likely
don't.

107 56 90 70 100 56
123 64 136 25 123 34


J

Jussi Karlgren, fil. lic. Jussi.Karlgrensics.se
Sw Inst of Comp Sc (SICS) Spr}kteknologi / Natural Language Processing
Box 1263, 164 28 Kista ph +46 8 752 15 00, fax +46 8 751 72 30
Stockholm, Sweden http://sics.se/~jussi/jussi-karlgren.html

Message 4: Re: 5.1229 Sum: Doctoral programs in Applied Linguistics

While Linguist may be an international list for linguistics, it is
strikingly parochial in many of its attitudes and discussion. It is
taken for granted that US terms of education etc are
universal. E.g. 'doctoral programs' are reserved in England for
students who are not good enough for proper Ph.D.s.; i.e. the meaning
of a program is quite different in the two educational systems. So
you're not going to get, say, UK universities answering such surveys
because they're not particularly proud of programs. Hence such a list
is going to make it appear that you have to go to the US to do a
Ph.D., when large numbers of UK universities would offer the
qualification, and many now have a program. I.e. the question about
Ph.D.s is culturally loaded and ethno-centric.
Vivian Cook