On Thu, 19 Nov 2015 20:39:03 -0500, Alex Fink <[log in to unmask]> wrote:
>If a purely-written language were around in the wild and had time to evolve, you'd expect this sort of thing to play a not insignificant role in the synchronic situation for what the apparent components were.  (The graphical analogue of breaking in sound change, I suppose.)

Yes, certainly. I can imagine that happening in UNLWS all the time, given the relatively unconstrained nature of the glyphs there (freer than, say, Chinese), although I suppose I'm unclear on how the components of an UNLWS glyph would be regarded, anyway. In U2D, the components are more or less evident from inspection, but even more rigid than Chinese due to the shape restrictions. I think that if the components were to evolve, very few of them would offer the possibility of eventually being seen as a combination of others. Though, perhaps that is not quite accurate; for one, a few glyphs already do look a bit like a combination of smaller ones (perhaps bad design on my part), and it might be that the current arrangement of components is too simplistic for native writers, such that it would eventually evolve to have far more than five possible shapes.

>The few examples of large UNLWS texts we've worked on, mostly relay entries, show the same sort of structure.  What wasn't clear to us is how much this is inherent to the structure of human-comprehensible narrative or naturally occurring constellations of participants, and how much of it could be reckoned an artifact of translation, a leftover from linear style.  If the latter, then, okay, perhaps we'd find it difficult, but we're hardly native speakers.

I find it hard to believe that native writers would tend towards less visual distinction, for clear communication. At least, it might be easier to read denser texts, but that's not particularly easy or natural for us, in the linear world, and yet we are native speakers there. I suppose what I'm suggesting is that, as linear-natives, it should be easier for us to read 2D spaghetti than a 2D-native, compared to our baselines. Because, I find, at least, when I read U2D, I tend to trace the lines quite deliberately, and don't gain much from the overall shape except a guide as to which line to follow next. Whereas I might expect a native to gain more at a quick glance, looking at the glyphs rather than their connections, and thus they might perhaps be more sensitive to spaghetti.

Though, of course, it would be much easier in absolute, being natives. And even if it's easier to read clearly, it might be much harder to write, such that most text is spaghetti regardless. And, I suppose, that's only considering ease of reading and writing. 

>The question of what UNLWS poetry is like continues to vex us, but this suggests a possibility: if an ordinarily text has this sort of clumping, i.e. has gaps at all scales, large and small, in its layout on the (large) page, then maybe a way to be poetic is to write a text that *doesn't*, that has a nice uniform-density mesh of referents and events at all scales.  (Without, of course, achieving the effect by just scrambing an ordinary text and leaving a spaghetti of lines everywhere.)
>What is U2D poetry like, do you know?

As far as poetry is aesthetic writing, I think I agree with your suggestion. The most aesthetic writing may be gridlike, with no pronouns or tangled connections (this is far easier to do in U2D than UNLWS, given the circular verbs and square nouns). Though, that might just be my personal preference. Perhaps along similar lines, any 'pure' geometric arrangement might be considered poetic; a perfect loop, or hexagonal arrangement, or something similar. Unfortunately, arranging glyphs so that they look like something else from afar doesn't work nearly as well in U2D as in UNLWS, because of the node-like shapes that dominate.

More linear-native poetic features might have analogues in 2D, such as repeating visual motifs or space. In my section of "The Raven", I tried to draw the "knocking" as onomatopoeic as possible, though that was mimicking linear poetry very deliberately. In a certain sense, I think onomatopoeia is possible here, through space and repetition like that (though, that might be an exceptionally obvious case), or with writing style, weight, and so on. In an abstract sense, the relative arrangements of glyphs beyond their bindings might be the most productive; this is, actually, so easy in small texts that it might hardly be considered poetic at all. You might be able to convey certain attitudes or feelings in how glyphs are drawn or connected: thick/thin relations, rising or sagging, stronger/weaker or larger/smaller glyphs, and so on.

Having said that, some 2D poetry may be visual art. Or, visual art with minimal bindings. I suppose that's not much to do with linear poetry, but it seems to be along the same lines, I think.

>Mm, what would the rule be, then?  A verb which is bound to by a collection of dot-relations, the one you call "Is (equivalence)", displays this reading? or something else of that sort?  Perhaps you could still analyse that construction as something like a periphrastic distributive, though that might be speaking uselessly loosely.

The dot-relation is just an explicit equivalence, actually not even asserting equivalence more strongly, just a form of the generic blank line that can be bound to. In that sense, I feel uncomfortable ascribing it special meaning when bound to verbs. As you say, that's periphrastic, which to me feels unnatural in the context of bindings. But perhaps that's what it would end up being, or at least, that as long as the subject/object relations were themselves being bound to, the verb could only be understood to be abstract. That, even, isn't quite correct, since you might want to ask whether or not someone did something, or who did something. In that case, you may bind other things to the verb that refute its abstractness, such as location. So in short, I don't know what the rule would be.

>OK, if U2D plurals are constrained to be collective like that, it indeed wouldn't help much.  But that's not a necessary restriction on plurals in familiar languages, or at least the unifying factor of the 'groups' can be as tenuous as you like.  If I gave you just the extension of, say, "former roommates of Nobel laureates", only a preternatural trivia maven could observe what makes them a 'group', but it's a perfectly cogent plural noun phrase.

I really mean a group more in the sense of action than selection. Of course you can construct any noun phrase. But the idea is that if you have nouns grouped together, their group is talking about them acting as a group, rather than distributing among them individually. It's true that spoken language doesn't have this distinction, but I don't think it's the same distinction as the one you're pointing out here.

>I think the UNLWS approach to "Which of X fulfils Y?" would be just to list all the elements of X and ask "Does X_i fulfil Y?" for each; since the yes-no question morph is relatively lightweight, that wouldn't be too distracting.  The same could also be done with a stack.  Either approach loses the implicature that there is at least one X that fulfils Y, but how often do we really need that implicature?

I think the more important implication is that there is exactly one X that fulfils Y, or at least that you are being asked to choose from a group, rather than check all that apply. That exclusivity, I think, is what is inexpressible otherwise. How useful it is is another matter; I think it's at least important for rendering speech. It's presumably also useful for talking about logical implication, even in the context of, say, a detective story.