At the time this thread began I'd been wanting to write some further thoughts I'd had on the artlang/engelang distinction; it had been sinking lower and lower down my stack-based to-do list, but the happy resuscitation of this thread restores it to the top of the stack.

Jörg Rhiemeier, On 22/05/2011 17:41:
> On Sun, 22 May 2011 14:54:45 +0100, R A Brown wrote:
>> My understanding is that
>> engelangs are "designed to specified *objective* criteria,
>> and modeled to meet those criteria."  Whereas artlangs will
>> have more *subjective* aims, e.g. aesthetically pleasing.
> This is also my understanding.  An engelang is distinguished
> by being designed according to a specified set of objective
> criteria, while an artlang is based on more subjective notions.
> This precisely is the difference between engineering and the
> fine arts.

I of course agree with Ray and Joerg: artlangs differ from engelangs as art differs from engineering; and the "designed to specified objective criteria" criterion is one I proposed when first drawing the distinction.

But I think there's more to be said about the distinction. As much discussion over the years has shown, the Objective Design Criteria Criterion seems not sufficient to characterize on its own the art/enge distinction. I don't have any clearly articulated additional criteria to add, but more can be said about the characteristics of the two (Art & Enge) prototypes. To attempt to boil them down to theObjective Design Criteria Criterion (as I myself did), is to fail to fully apprehend the distinction.

The Art approach:
* sees language as a natural and/or organically evolved entity, in which parole is as important as langue
* is akin to the impulse to make miniature models of the natural or cultural world -- to subcreate versions of reality
* is comparable to the invention of physical geographies, or imaginary actually inhabited towns
* creates a fictional entity

The Enge approach
* sees language as an abstract system, a machine, a set of rules, pure langue independent of any parole
* is akin to the impulse to invent solutions to problems, to devise functional improvements to existing design solutions
* is comparable to the design of hypothetical city plans and buildings (i.e. not descriptions of how imaginary cities and buildings happen to be in the world they exist in, but rather designs for Ideal solutions for the functions cities and buildings have)
* is comparable to the design of legal and electoral systems, of laws and voting methods, (again, Ideal solutions rather than simply the modelling of ones that happen to exist in an imaginary world, even if none of the rest of the imaginary world has been subcreated)
* creates a nonfictional entity

These aren't neat criteria than one can easily apply to classify conlangs. But they do characterize two very different approaches to invention, and two very different mindsets that conlangers have. Just as linguistics more than almost any other discipline is at the crossroads of many other heterogeneous disciplines, so the invention of conlangs seems, more than the invention of almost any other sort of thing, to straddle these two very different kinds of invention.