Re: Language comparison

RB:

> The only problem is that any simple or primitive hominid languages became
> extinct many, many, many millennia ago, way before any method of recording
> them. Without time-travel we shall never know them.
I meant "primitive" as in those used by modern-day African tribes, and
as some people would treat e.g. "Black English Vernacular".

> If it were required, any existing human language can develop itself
> (principally by expanding vocabulary, as English has had to) in order to
> express the profoundest of thoughts and certainly all the advances of modern
> science and technology.
... which is what I would call "minimal sufficiency". Please assume
that everyone has the same fluency in the language(s) involved.

I am *not* trying to say that any full language is *insufficient*.

An analagous thing would be (I hope I linked to the paper that
describes this point) that some computer-language features can be
implemented merely by adding new "vocabulary" (library functions),
whereas others (e.g. recursion, lexical scoping) would require the
language *with* it to be emulated to get the same power.

I am only talking about the latter.

> Better at what? As Chris Bates so rightly observes: "Without criteria the
> entire question means nothing whatsoever."
> As for conlanging - there are several different types, with *quite
> different* aims. To treat them all the same is very misleading.
I did not intend to do so, only to say that some languages can better
fulfill a given goal - even if both do take on that goal. If one does
and the other doesn't, than it's an invalid comparison, of course. The
choice of goals is, in all probability, arbitrary.

> We on this list like to help one another with ideas, whatever the type of
> conlang, and we respect our very different aims; we do not try to judge
> which conlang (or natlang) is 'better' than another. (Because in practice
> auxlangers do like to compare one language with another for superiority,
> they have their own list)
*see above* Not trying to compare *goals* as being better or worse -
they're probably axiomatic - just implementations *towards* those
goals.

> For what purpose? That is a serious question. Until you know what the
> objectives of your conlang are, you will not know what features of
> natlangs are relevant in any case.
I can answer this one less theoretically. For me-personally, my goals
could be summarized as maximum communication - compact, dense,
multimodal,


TM:

> So don't use a technical term ('Turing complete') when you don't mean
what it means ;)

How is my use of "minimum sufficiency" different from that?

> Err---no. These would be questions used to help me make a decision. At
the moment, all I see is an assertion that not all fully fleshed-out,
complete languages are equal.

I see your point; my intention had only been to present this as a
theoretical possibility for starters.

I should clarify (again) something that I think you mistook about what
I was saying; I am not - necessarily - saying that there exist natural
languages that are better than each other (though this is a
possibility). I would be curious to see if there are, but my guess is
that they are mostly equal. This doesn't mean that it must necessarily
be so for any language, though.

> (If you don't yet hold this view but are only looking at it
academically, why not throw a few reasons into the ring for us to poke
at and prod, so that maybe the discussion can have a bit of life.)

Essentially, that's what I was doing. So, okay, I'll try [below].
However, I should say that my *guess* is that describing these
potentially-differential features is most of the work of them; I don't
necessarily expect to be able to come up with good enough inventions.

> Depends on the thinker. My consciousness (or the bits of my thought I
have access enough to to communicate) are serial enough and in the
English language first and foremost, though obviously irrelevancies pop
up and when writing emails I'll suddenly leave a sentence and go write
another one elsewhere. But I have no access to the process which brings
up these thoughts, so a written medium like email is adequate enough
for the communication of my ideas.

*nod* Might it not be (S-W style) that the way you think consciously
is serial because your language is also? I recognize that this might
in fact *not* be true, and if so, nonserial language wouldn't matter
to you. I know that for me, it feels like a significant shift to
translate thoughts to serial language - even for talking to myself
(something I don't do that often) - and I feel I loose something in
this. Hence why I would want to experiment with finding modes that are
"better translations" of my actual thoughts.

(It's not so much that my thoughts are not *directed* - e.g. skipping
around - that's still serial. It's that they are interconnectedly
parallel or more; I don't think any serial description would be
adequate unless it chose to highlight one particular subset of the
actual thought process. Asynchronous multithreading might be
analogical.)

> As it stands it most certainly is not---we don't have a definition of
'better' yet and the best we can do is badger you until you give us one

I thought I had explained; I guess not. "Better", then, as I use it,
implies that there is an implied standard - a goal or set of goals -
and that it is possible to measure how well some particular
implementation (or language) meets those goals. Yes, some goals will
be qualitative - e.g., aesthetics - and some will probably not be
bounded measures - e.g., "power" in a sense somewhat (not 1-1)
analogical to the CS one in the article I linked to. Some will,
however - e.g., writing systems that are legible from any angle or
direction of reading.

> I've yet to hear of a language that cannot adequately express the fact
that yesterday, I ate a bucket of stew. IOW, no language lacks the
features necessary to discuss events occurring in past time. Perhaps
people don't 'compare features' because all languages are equivalent?
Maybe this is why the consensus is that all languages are equal?

I'd disagree. What you are implying is that the current feature set of
most/all languages is the maximal one. To me, that seems like an
arbitrary and somewhat cynical perspective, since it places you in a
position of assuming that it is impossible to improve. Whether it is
or not, you won't ever discover an improvement.

(So that all languages implement "past time description" does not
really address this...)

> My gosh! We've gone from a hypothesis to a theory without doing
anything!

My apologies for using the language in anything but the most precise
possible way.

> the idea seemed interesting and pre-contempory, both of which
are currently interesting to me.

"Pre-contempory"?

> And it is my understanding that this is precisely how you get yourself
a PhD, if you succeed (not that I have one, mere undergrad that I am).

Interesting, since I, a fellow undergrad, have been considering
writing a thesis exploring something related to this very topic.

> So tell us: What questions *do* answer yours? I'm satisfied by what
leaves you wanting, and I don't know where to turn next.

A solid refutation would be a proof of what I pointed out above as
your assumption that the current feature set is the maximal one. Or,
in fact, a meta-proof of whether such a proof is even possible. I'll
grant that you need not try to proove *my* point, but you are making
no less of a handwavy support for yours than you claim I am.

> To me, this seems to mean you have 'better by measure X'.
Yes, it does.

> Today we already have that; if we are trying to communicate with me, English is
*by far* your best choice,

As I said, you need to assume that everyone has equal fluency in all
available options.

If, in fact, the decision of people with equal fluency among languages
(with equal social status, etc., for outside factors) for what
language to speak in any given situation is completely arbitrary or a
matter of idiosyncratic aesthetics, then sure, that would be a partial
refutation of what I've said. However, that would only say that the
languages *they know* are fully equal; it wouldn't have any way of
addressing languages that they might theoretically learn afterwards.

Much like I could learn various BASIC dialects, and not care which I
used... until I learned one able to do recursion. I'm looking to find
that next new feature - the one I don't yet know.

So, since you asked - some (off the top of my head) guesses at what
might constitute such a feature. (If you want to discuss them, I'll be
glad to, but I think that the merit of any particular guess is not so
much at issue here as whether there might *exist* some such feature of
sufficient merit to be considered a significant upgrade...)

Some of these are qualitative, some quantitative, some boolean. No
particular order.


* little effort to use
* semantically dense (i.e., minimal wasted space [bounded by necessary
redundancy)
* multimodal - maximally using whatever means of information transfer available
* as clear/ambiguous as desired
* corollary: not overspecified (e.g., ubiquitous gender when wasteful
of space or otherwise undesired)
* runtime encryption (if wanted)
* writing system legible from any direction / inversion (e.g. flipped,
or on other side of a window)
* writing system that fully uses its medium (e.g., non-serial use of 2d space)
* minimal loss in translation from thoughts, and minimum possible
potential for confusion between users
* usable in any given environment/situation (e.g., underwater, loud
rooms, with hands full, with mouth full, etc.)
* possible to carry multiple meanings (if desired)
* advanced version: carry multiple meanings, some of which are only
understandable if you have a different mode of access (e.g., a more
advanced version of shaking your head when you're talking, such that
hearing-only listeners don't get the "this is false" message)

...

I could go on, really, but I don't think that's necessary here.
Hopefully I've given some idea of the sort of thing I mean.

 - Sai

P.S. No need for the snarky tone; I was not rude to you.

Replies