Hallo conlangers!

On 21/03/2018 18:30, Raymond Brown wrote:

> On 21/03/2018 16:36, Jörg Rhiemeier wrote:
>> Hallo conlangers!
>>
>> On 20/03/2018 01:14, Sai wrote:
>>
>>> FWIW: https://s.ai/essays/on_the_design_of_an_ideal_language
>>
>> A very sophisticated and ambitious list of principles which sets the
>> bar *very* high, and partly conflicting with each other: while
>> principles #3 and #8 call for very little redundancy, principles #9
>> and #10 call for a good deal of redundancy. How do you get something
>> noise-resistant and error-proof if every possible string of phonemes
>> (or whatever) is a valid utterance?
>>
>> As you certainly know, I do not think that an "ideal" or "perfect"
>> language is attainable at all,
>
> I don't think "ideal" and "perfect" are synonymous, unless one is using
> "ideal" in a purely Platonic sense, meaning 'of or pertaining to the
> Perfect Form (ἰδέα).

Sure. I for sure am not a Platonist, but whom am I telling that, you of
course already know. I don't know whether Sai is a Platonist or not,
though. What I have noticed in Sai, though, is that the is not
interested in naturalistic conlangs but wants to create a language that
is some sense superior over natural languages.

> One meaning of _ideal_ is "satisfying one's conception of what is
> perfect or most suitable", i.e. it's subjective, not objective.  This
> seems to be the way Sai is using the word, since he writes: "If you are
> wondering what exactly I mean by "ideal" ... take a look at this ..."

Yes, "ideal" is a more subjective notion than "perfect", and does not
necessarily involve Platonic ideas of an ideal Form.

> He also acknowledges what you observe in your first paragraph above,
> since he writes: "Also to be clear, these principles are intentionally
> in conflict with each other. One cannot fulfill all of them; one can
> only choose a preferred balance, and optimize for that."

Fine - it is certainly a trade-off between conflicting design goals,
where one can perhaps find a subjectively "ideal" balance but no
absolute perfection.

> I don't have a problem with someone having a _subjective_ set of ideals
> and knowing that they can't all be achieved and that there will have to
> be a trade off between them.  I think that per_se is an admission tht
> the resukt will not be perfect.

Right - in a sense, Old Albic is perhaps indeed an attempt at an "ideal"
language - one that fits my personal notion of the Elves of prehistoric
Britain best. and certainly, that language is as imperfect as any other
human language.

> But having written that, I'm not sure that Sai's 1st guiding principle
> "Principle of Good Representation" is achievable.
>
> To me to say something is perfect is making an _objective_ statement: it
> is _absolutely_ complete and cannot be further refined.

I understand - and I strongly feel that perfection is unattainable. A
best solution of a given problem, in contrast, *is* attainable, of
course - even if such a judgment requires a set of criteria (and
priorities where the criteria conflict) which is usually not objectively
definable.

>> and I do not endeavour to create something like that in my conlanging
>> projects.
>
> The notion of "the perfect language" is IMO complete nonsense. But
> setting out subjecive ideals a person might aspire too and realize that
> not all can be achhieved and that there will have to be trade off
> between these 'ideals' seems tolerable to me, especially for an engelang.

I concur with this. there is no such thing as a perfect language, though
a conlang can be more appropriate to some kind of "purpose" or design
goal than others. For instance, I feel that Old Albic is more
appropriate to its "purpose" - the portrayal of a language hat could
have been spoken in prehistoric Britain by a civilization imagined to
believe in the same ideas about the human condition and the meaning of
life as I entertain - than, for instance, an oligosynthetic language
like my dormant Quetch project would be. But Quetch has been designed
for a very different purpose - to test the idea that a "speedtalk" in
the Heinleinian sense, i.e. an oligosynthethic language in which all
morphemes are exactly one segment long, actually leads to shorter
utterances. I haven't yet done such a test; I haven't even developed the
language sufficiently. But I expect the brevity of the morphemes to be
cancelled out to a substantial degree by the need of many multi-member
compounds to express basic notions natlangs (or naturalistic conlangs)
have root words for. Of course, I couldn't perform such a test with
something like Old Albic.

>> Most of my conlangs are *naturalistic* languages, designed to
>> resemble plausible natlangs, guided by what fits the setting chosen
>> for them, and my personal preferences.
>
> Exactly.  Britainese will not be perfect nor ideal.

It won't be - but it may be an aesthetically pleasing and historically
plausible answer on the question "What would the Romance language of
Britain be like if there had been one?". And I feel that it is promising
to do a good job of this.

>> I have had ideas for experimental languages, but none of them with a
>> claim to perfection, either, and all of them are now either dead or
>> dormant.

Which shows that these projects did not satisfy me - I either found no
way to make these things work out, or simply lost interest in them.

--
... brought to you by the Weeping Elf
http://www.joerg-rhiemeier.de/Conlang/index.html
"Bêsel asa Éam, a Éam atha cvanthal a cvanth atha Éamal." - SiM 1:1