LINGUIST List 2.368

Sunday, 28 July 1991

Qs: 40's radio, Ergative, IBM or Mac, English

Editor for this issue: <>

Directory

1. Randy Allen Harris, Query: HLS's radio program
2. Alexis Manaster Ramer, The origin of the term "ergative"
3. , IBM or Mac??
4. stan kulikowski ii, distinctly english algorithm?

I'm looking for some information about Henry Lee Smith's
radio program from the forties. Virtually anything will
help ( all I know at the moment is that it existed, that it was
fairly popular, and that it concerned, at least partially,
American dialects), but specifically does anyone know
(1) its name, (2) any particulars about its content and/or
popularity, (3) if any tapes survive, and (4), if the tapes are
available, where?
Thanks.

I would appreciate any information regarding the origin of
the term "ergative" for an article I am preparing.

I need your help. I am a doctoral student in New Testament and Greek and am
writing a dissertation on linguistic method for studying the book of Revelation
and apocalyptic literature. I hope to teach and/or work in text translation.
Given these interests and needs, would you recommend the IBM or Macintosh, and
which machine would you suggest from either category? My main concerns are:
1. word processing and database applications suitable for multilingual work,
2. educational software for my kids (ages 6 and 2.5). As far as the kids are
concerned, is color essential? The Mac has superb screen graphics, which seem
to me to make educational software quite appealing, even without color.
Your advice is greatly appreciated.
Steve Waechter
Mid-America Baptist Theological Seminary
SWAECHTERUTMEM2 (BITNET)

 i have been collecting network sources of text in languages other than
english, we have been collecting several samples of network text in various
languages so we can test some software which measures textual complexity. here
are some of the sample sizes to date.
 turkish 7498168 bytes
 polish 2342845 bytes
 spanish 1190638 bytes
 esperanto 1075657 bytes
 french 907609 bytes
 gaelic 888091 bytes
 croatian 403917 bytes
 german 390054 bytes
 in just about all network language sources of text, english is a fairly
common contaminant. i would to screen samples like these for the presence of
substantial inclusions of english. i am not concerned with borrowed words or
phrases in which writers are using crossover terminology, like 'software' which
appears commonly in spanish network text. what i want to identify is portions
of text (a couple sentences or more) where english has been quoted from other
other writers, often brought in by cut-and-paste file transfers.
 does anyone know of an algorithm which can identify english text as distinct
from most other languages?
 here is what comes to mind... search for string occurrences of common
uninflected english words like
 ' of '
 ' with '
 ' and '
 ' it '
 ' this '
 but which of the most common words are most distinct for english?
 are there other approaches, such as syllable or character frequency searches
which could help distinguish english within a few lines of text?
 eventually i am going to want to generalize the algorithm to a wider
application, distinguishing file inclusions from other languages. once in a
while we get a sample from an unidentified source and need to tell what
language it is... the students i work with have developed a little game which
produces a maybe-algorithm that says
 'of' + 'with' + 'the' =?= english
 'di' + 'per' + 'il' =?= italian
 'de' + 'con' + 'el' =?= spanish
 'auf' + 'mit' + 'das' =?= german
 'du' + 'por' + 'le' =?= french
 'od' + 'u' + 'iz' =?= croatian
 'ag' + 'agus' + 'ach' =?= gaelic
 'ke' + 'pri' + 'mi' =?= esperanto
 'ze' + 'nie' + 'sie' =?= polish
 'bu' + 'bir' =?= turkish
 if any of you know of this kind of work that has been done, or have
suggestions what to try, i would appreciate any help.
 stan
 . stankuliUWF.bitnet
 ===
 | | close your eyes, my darling, or three of them at least
 --- -- old venusian lullaby