Re: Throwing out the tree-structured grammar (SF Xenolinguistics FAQ)?

On Friday, June 10, 2005, at 06:21 , Keith Gaughan wrote:

> Steven Williams wrote:
>
>>> Architecture: seriously alien grammars may throw
>>> out the whole terrestrial "tree-structure" scheme,
>>> replacing it with some bizarre kind of "stack"
>>> or "hash", but I feel no urge to attempt to
>>> describe such horrors.
>>
>> How would one do that?
>
> Well, the stack thing's all already been done: take a look at Fith[1]
> for an example.
True - and we had some discussion on stack-based langs a few months back.
IIRC I made some proposal for a conlang to be even more 'stacky' than Fith
    :)
=============================================

On Friday, June 10, 2005, at 07:26 , Patrick Littell wrote:
>
> To be needlessly picky about it, a stack language -- that is, one whose
> underlying computational implementation is a stack -- is still a
> tree-structured language.  More precisely, it's the reverse Polish
> notation (postfix traversal) of a tree.  If it weren't, the stack would
> either underrun or (eventually) overflow.
I don't think you're being needlessly picky. What you say is quite correct.
  IMHO the author of the FAQ that Steven quoted is being rather vague, and
"I feel no urge to attempt to describe such horrors" seems a bit of a
let-out to me. Has the author thought through what s/he's writing? Without
an attempt to describe 'such horrors' I take the FAQ with a big pinch of
salt.

It would certainly be a very bizarre kind of stack indeed that is not
'tree-able' - indeed, it would, as Patrick says, underrun or overflow.
=============================================

On Saturday, June 11, 2005, at 12:13 , Henrik Theiling wrote:
[snip]
> I was picky in the same way here a while ago, but a *true* tree or
> stack structure still means the structure is sufficiently different
> from human languages.  E.g. the stack modification operations of Fith
> are really funny and non-human. :-)
Also true - but Fith was inspired by the programming language Forth. So
while Fith is sufficiently different from human natlangs, it is similarly
to a human devised programming language.

But you are right to draw attention to a *true* tree structure. What
precisely does the author of the FAQ mean by 'terrestrial "tree-structure"
  scheme'?
==============================================

On Saturday, June 11, 2005, at 03:54 , Paul Bennett wrote:
[snip]
> I cannot see that it is possible to have anything recognisable as language
> that is not reducible (or morphable, or pick-your-verb-able) to some kind
> of tree. The universe of perception is things happening to objects,
> relationships between objects, and relationships between events. That's
> pretty irremovable from a tree-able structure, isn't it?
It certainly is. If our alien is not dealing with things happening to
objects, relationships between objects, and relationships between events,
what in Cosmos is he/she/it trying to communicate?
==============================================

No one has mentioned the FAQ author's bizarre "hash". While this may,
properly executed, be a convenient way for a machine to store and retrieve
data quickly without (in theory) any searching, how would it contribute to
direct communication?

For those unfamiliar with hasing, I quote from the British Computer
Society's Glosary of Computing Terms:
"Hashing is the process of calculating a numeric value from one or more
data items. While this value obviously depends on the value of the data
item(s), it does not depend on the meaning attached to them. It can be
used ........."
"It is also an important technique in storing values in a data structure
known as a _hash table_. The calculated value can be used to mark the
position in the table where the data item should be found, enabling it to
be accessed directly, rather than forcing a sequential search."
"A _collision_ occurs when two data items give the same value under
hashing, and should therefore be stored in the same place in a hash table.
"

A 'perfect' hashing algorithm in theory produces no collisions; in
practice it is difficult to avoid them entirely and some collision
handling procedure is needed.

So a race of aliens has evolved that performs near perfect hashing
algorithms on the morphable items, which presumably can be numerically
evaluated by the aliens? Presumably alien A passes the resultant hash
table to alien B. Very nice - but how in Csmos does alien B know what
order to retrieve the morphable items? It seems to me that with the hash
table, alien A must also pass a string of numeric value so that alien B
can apply each value in the string to the hash table and extract the
morphable items in their correct order.

Bizarre!!!

Maybe that is why, although there have been stack-based programming
languages, there is not (as far as I know) a hash-based programming
language     :)

My impression is that the author of the FAQ has not really thought through
what s/he has written. IMO without the attempt to describe 'such horrors'
it does not really say anything more than "Alien grammars may be quite
different from any grammars of our familiar earth languages."

So what? Until we encounter any such aliens, we will not know.

Ray
===============================================
http://home.freeuk.com/ray.brown
ray.brown@freeuk.com
===============================================
"A mind which thinks at its own expense will always
interfere with language."         J.G. Hamann, 1760

Replies