Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!linus!philabs!cmcl2!seismo!think!mit-eddie!genrad!decvax!cca!g-rh
From: g-rh@cca.UUCP (Richard Harter)
Newsgroups: talk.philosophy.misc
Subject: Re: The meaning of language (information content)
Message-ID: <10513@cca.UUCP>
Date: Sun, 12-Oct-86 05:02:05 EDT
Article-I.D.: cca.10513
Posted: Sun Oct 12 05:02:05 1986
Date-Received: Thu, 16-Oct-86 07:32:30 EDT
Reply-To: g-rh@cca.UUCP (Richard Harter)
Organization: Computer Corp. of America, Cambridge
Lines: 78
Summary: Decryption is not translation

In article <> tedrick@ernie.Berkeley.EDU.UUCP (Tom Tedrick) writes:
>>>	My simple point was that, for example, a language created by
>>>some intelligent creatures who have all died out would become a
>>>"meaningless" set of symbols (assuming that language consists of
>>>strings of symbols or something of that nature.)  If humans were to
>>>try to interpret that language, they would only have the barest chance.
>
>>   This is what you believe.  However, since you offer no evidence or
>>justification for this opinion, why do you expect others to accept it?
>>Exactly the opposite seems true to me; 
>
>Yes, this is a point which requires some technical knowledge
>to understand.
>
>There is bound to be some "information" (in the information-
>theoretic sense) in the samples of the language that remain.
>Cryptographers have frequently managed to break what were
>thought to be impossibly difficult encoding schemes. There
>are some very clever fellows in the "Cryptography Cabal" ...

	Well there would certainly be "information", but that does
not necessarily suffice for translation.  Text in a natural language
has an information signature; encryption is basically a process of
disguising that signature without destroying the "information".
The effect of encryption is to transform the information signature;
cryptanalysis ultimately works by deducing the transformation
process by comparing the transformed signature with the original
signature.  It does not, however, reconstruct the original
"meaning" which is a different issue.

	In practice the cryptanalyst working with a foreign language
[i.e. one she does not know] can reconstruct the original
plain text; however once she has reached that point she has
to turn it over to a translater.

	The difficulties of translating text in an unknown
language de novo are enormous.  My recollection is that it has
been done in a couple of instances [rather incompletely] and that
most dead languages for which there are no cross translation
aids simply remain untranslated.  In any case, the problem is
how do you interpret the symbols that you have.  In the story
"Omnilingual", H. Beam Piper made the point that scientific
knowledge would likely be discernable from context because of
the universality of scientific knowledge.  For example, suppose
we found a chart that was recognizably the periodic table;
that would give us the names of the elements.  Similarly numbers
would probably be deducible.  This method has actually been
used, e.g. the Mayan astronomical tables.  However it clearly
has its limits.

	The nature of the difficulties can be imagined by
supposing that you are given a number of books [without pictures]
in an unknown language and are given the task of reconstructing
the language.  Without knowledge of the linguistic principles
of the language you don't even know what kinds of parts of
speech are being used.  [Some languages don't have the noun-verb-
adverb-adjective-preposition categories that we are accustomed to.]

	Even if languages are known, translation has its problems.
Quite often the original text rests on cultural presuppositions
that must be reproduced and included if the translation is to be
at all faithful.  For example, it is almost impossible to make a
translation of a work of oriental philosophy that is even compre-
hensible unless you have some background in oriental philosophy.

	So, as far as "evidence" is concerned, it is simply that
the success rate of reconstructing human written language in the
absence of translation keys is very low; we can only expect that
it would be worse for our hypothetical aliens.  We observe further
that translations difficulties arise even with humans because of
cultural difficulties.  We can only expect that an alien culture,
driven by a different physiology and different emotional constants,
would be even more difficult to deal with.

-- 

Richard Harter, SMDS Inc. [Disclaimers not permitted by company policy.]
	For Cheryl :-)