Formal vs. natural languages (was Re: Oligosynthetic languages in nature.)

Hallo!

On Sun, 29 Mar 2009 18:18:29 +0100, R A Brown wrote:

> Andrii Zvorygin wrote:
> [...]
> > Wonderful, thank you so much for your great knowledge. It is wonderful
> > that you have shown the similarity between for instance the English
> > alphabet of 24 letters and that of the Amino Acids with 20 varieties.
>
> 26 letters are currently used in the English version of the Roman
> alphabet, methinks.
>
> But there seems to me to be some confusion here between 'formal
> language' and 'natural language'.
Seems to me, too.  And a misconception of the meaning of
"natural language".

> A formal language is any language generated by a _formal grammar_. A
> formal grammar is a "fully explicit device which specifies, for a given
> initial set of elements (the 'vocabulary' or 'alphabet'), the complete
> set of strings of those elements which are in the language defined by
> the grammar." [Trask]
>
> Possibly the DNA code does constitute a formal language - I don't have
> the competence to say whether this is so or not.
It does indeed constitute a formal language, as does the set
of proteins occuring in living organisms, with the set of
amino acids as "alphabet".

>       Certainly there is no
> doubt one can define an 'oligosynthetic' _formal_ language.
Yes.  Formal languages have little to do with human languages,
if anything at all.  For one point, the definition of a formal
language does not say anything at all about *meanings*.  Also,
while formal language theory has found a useful application in
programming languages, trying to treat human languages that way
has turned out to be less than successful.

> But the term 'natural language', as used in linguistics, means quite
> specifically: "A language which is, or once was, the mother tongue of a
> group of human beings."
Andrii's misunderstanding, it seems to me, lies in part in
miscomprehending the term "natural language" as 'formal language
that manifests in nature (i.e., outside the human cultural
sphere)'.  Which is NOT AT ALL what the term "natural language"
means.

>       AFAIK oligosynthesis is not found in any natural
> language in the linguistic sense.
Right.  Human languages, in order to have full expressive power,
need at least one *open* class of lexemes, i.e. one to which new
lexemes can be added whenever needed.  And that is precisely
what an oligosynthetic language is not: "oligosynthetic" means
that the set of lexemes is *closed*, i.e. no new lexemes can
be added to it.  Reality (and a fortiori, the vast universe of
things imaginable) is too complex to pigeonhole it into such a
closed list of "semantic primes", and that is the precise reason
why oligosynthetic languages are impractical and do not occur
among human natural languages.

> That does not, of course, mean that one cannot attempt an oligosynthetic
> _conlang_ - but so far attempts to do this do not seem to have met with
> success.
Precisely.  For the reasons I stated above.  Learning from the
past in this case means accepting that it just won't work.

> [snip]
>
> > Yes it is wonderful how cohesive we with you are. As the concept "As
> > above, So below" of the gnostics, and "Macrocosm, Microcosm" of the
> > Elizabetheans, the "As in area, So in point" of Geometry.
>
> Maybe senility is now setting in, but I just do not understand what this
> is about. Is it possible to rephrase this in a way that an old timer can
> understand?
>
> Thanks.
I think it is not senility's fault that you don't understand it.
I am young enough to be your son, and I don't understand what
Andrii said, either.

... brought to you by the Weeping Elf

Reply