Theiling Online    Sitemap    Conlang Mailing List HQ   

OFFLIST: Re: Non-linear full-2d writing (again)

From:Tom Chappell <tomhchappell@...>
Date:Wednesday, February 1, 2006, 1:08
--- Sai Emrys <sai@...> wrote:

> Tom - > > > I thought Jeff W's G.A. was a good approximation > to a foundation for an > > example of what you were looking for. > > > > I didn't think by any means all examples of what > you were looking for > > needed to resemble Jeff's much at all, but I > thought his is the best- > > worked example that's available on the Web. > > > > If, in fact, his G.A. is something "fundamentally > different" than what > > you were looking for, I must have misunderstood > what you were looking > > for. > > > > Could you explain, better, what you are looking > for, so I won't make a > > similar mistake again? > > Sure; I'll try to describe it as a diff, since GA > does have some > substantially similar features. > > (Also clause this by saying that while I've read > over the GA page and > think I understand how it works, I may have > misunderstood it.) > > First, GA is grid-based. In that sense, it is indeed > "two-dimensional" > - literally. It works as an array. My conception of > 2d is not as an > array, but freeform; something much more like > Ouwiyaru in that sense. > (Actually, Ouwiyaru I think is the closest example > I've seen to what > I'd like.)
Ouwiyaru is hard to find. I found it at
> Because of this, GA drastically limits its > connectivity. It only > allows immediate local connections - not the > (nearly) any-to-any > connectiveness that I would want.*
Ah. I think I see. In my post " Message number 131804 From: tomhchappell <tomhchappell@...> Date: Sat Jun 4, 2005 3:49 am Subject: Re: Non-linear / full-2d writing systems?
>From tomhchappell@... Fri Jun 03 21:34:47 2005
Message-ID: <d7r8c5+6062@...> Date: Sat, 4 Jun 2005 03:49:25 -0000 Reply-To: Constructed Languages List <CONLANG@...> To: CONLANG@... In-Reply-To: <Pine.WNT.4.63.0505310936190.14940@acer> " I mentioned that subordinate clauses might have long-distance co-references or long-distance dependencies; graphically showing this in a NLF2DWS might introduce loops. The same might also be true of a two-clause sentence with an cataphor in the first clause co-refering to a noun in the second clause, while an anaphor in the second clause co-refers to a noun in the first clause. This type of sentence has a name; but it escapes my mind, as do the best examples. Here is an example I've made up; If that man knows her, he loves Katy.
> Second, there do not seem to be significant > differences (other than > the connectives) between the 'symbols' in GA. By > symbol here, I mean > the ones that are usable as characters - not the > funadmental parts, > which I'd think of more like 'radicals' in > Chinese/Japanese (and which > in those are indeed psuedo-2d, within their > characters). This seems a > suboptimal design; I would like the very shape of > characters to play a > *major* role in its visual and grammatical > structure, and indeed in > how it can connect.
But, how crucial should this point be? Is it really fair to rule out systems such as Jeff W's Glyphica Arcana just because they don't live up to this second point as much as you'd like? I think this is a point I would call "aesthetic", as you called the lines-mustn't-cross restriction I had referred to when talking about flow-charting.
> This point bears a little elaboration. In linear > writing, the actual > content of symbols is more or less irrelevant; they > can be whatever > shape you like, since all you care about is that > they are symbol > #12345 then symbol #12347 then etc. (Imagining that > each was > serialized, as is in fact done with hanzi/kanji.) > Their internal form > has nothing to do with anything. And, aside from > some small > exceptions, they are all moreorless the same size > (or treated as if > they were). > > In a 2d system, that would be a *huge* waste. > Symbols or meta-symbols > that have low content should also be small and > simple - e.g. a > pluralization 'morpheme' or a causal-connection one.
Ouwiyaru's symbols actually represent phonemes, or perhaps syllables, AFAICT. Maybe some represent fundamental, frequent function-particles; I haven't read it well enough.
> And I'd like that > symbols that are meant to connect to each other in > particular ways > show that.
Ouwiyaru doesn't satisfy that desideratum.
> > As a really really really simplified example: > > X-( *-Y > > Pretend that X-( is one character (the rightside is > an actual chunk of > it as written) and *-Y is also. (I just got out of a > neuroscience > class, so this may be a bit familiar in origin to > some of you...) > > The two connect naturally. Putting two X-figures > together (or Ys) > simply wouldn't result in a connection, because they > don't mate > appropriately. > > That's a non-overlapping example. Ouwiyaru has what > I think are some > pretty neat examples of overlapping connections - > not as extensive as > I'd like, but definitely of the kind that I'm > talking about.
I didn't catch that on any of my (admittedly non-numerous) readings of Ouwiyaru, yet. Would you mind giving me an example?
> Also, that's an example that has no semantics in it, > which is an > excessive oversimplification. I would like to see > the subcomponents, > the very graphical form, be semantic and otherwise > functional. > > My 'sketch' paper - > - is a > pre-alpha version of what I mean in this respect - > see figures 1-8, > the 'commercial transaction' frame - > > Note how the form > indexes the semantic, even in its subparts [like the > 'goods' node > travelling within the character to indicate > ownership]. It would be > better if the *form* was also somewhat ideographic - > i.e. looked > vaguely like what it is intended to represent - but > its *function* is > what I'm referring to here. (See the text of the > paper for an > explanation, since the diagrams aren't completely > self-explaining.)
No, I'm sorry, I don't get it. To me, both your sketch and Ouwiyaru suffer from the same deficiency that Carsten and I have complained about to Jefferson Wilson concerning Glyphica Arcana. Namely: 1) There is no explicit way to tell what the semantics are of a given way of assembling the elements, nor of the semantics of the elements within given assemblages. 2) There is no explicit way to tell how, given a certain semantic notion-or-idea-or-concept, to choose which elements to assemble, nor how to connect them, to express that concept.
> There are some other points I could add, but I think > those two are the > major ones that come to mind. > > > Jeff's _Glyphs_ are not all the same size nor all > the same shape. > > > > The "_symbols_" out of which the glyphs are > constructed are all the > > same size and all the same shape. > > See above for what I intended by 'symbol'. You I > think are talking of > what I called 'radicals'
Yes, the things Jeff uses for elementary symbols could be similar somewhat in use to "radicals" and "determiners". What I had in mind to compare them with were the "jamo" in Hangul. I think they are comparable to the things out of which the "jamo" are constructed. A glyph is comparable to a "syllabic block" in Hangul. But since neither the symbols nor the markers have phonological nor phonetic values specified by Jeff, and the symbols also have no semantic values specified by Jeff, they do _not_ have _precisely_ the functions that radicals and determiners have in Chinese and Akkadian and so on, nor for that matter that jamo or the marks out of which jamo are composed in Korean.
> - they aren't accessible to > the larger > (between-'word') syntax, so they're not very > relevant to it.
True enough, I suppose; except, perhaps, when (if?) he gets around to it, agreement and concordance and cross-referencing and flagging and indexing.
> > His symbols symbolize something like "distinctive > features". His > > glyphs symbolize something like compound-complex > words and/or simple- > > to-medium phrases. > > *nod* Though 'simple' seems to have a really low bar > (viz. the symbol > representing God, or Explore, or etc - while these > do in fact > represent very complex meanings, I think it a major > mistake from what > I know of category representation, frames, etc., to > try to 'write out' > these meaning rather than merely point to them.
Well, at least, it looks like you could tell what I meant. I think I see what you meant, too, though I don't think I see why you said it. Not that I necessarily disagree; I just don't know what convinced you, nor why I should agree.
> (I'd point out though that the latter is IMO a > separate question, not > necessarily related to 2d writing itself. An > interesting one, though.) > > > He hadn't, the last time I looked, explained how > to derive the > > meaning of a glyph from the meanings of the > "elementary symbols" of > > which it was constructed, and their relationships; > nor, given an idea > > or concept, how to construct the "glyph" to > represent it. (No > > obvious connection, in other words, between a > glyph's visual > > structure and its semantics.) That's why I > thought his G.A. was a > > good _foundation_ for the sort of NLF2DWS you were > looking for, as > > opposed to being a full-fledged _example_ of such. > It's still IME > > the best-worked-out example available on the web; > if you know of a > > better one, or even one almost as good, please let > me know how to > > look at it. > > I think he was using the same 'derivation' rules as > e.g. Leibniz - > kludged ones. This is another supporting point for > my statement above > - like it or not, he's indexing something he's not > really writing out. > Symbols are arbitrary; I think (but am not certain) > that this is an > inevitability. Having them be *translucent*
A no-longer-really-a-joke term meaning "Not quite transparent, but also not opaque"?
> - that > is, have some > morphemic or sublexemic in any case regularities > that make them more > memorable or guessable or whatnot is a Very Good > Thing. But trying to > have them as a general rule be *derivable* is, in > all attempts I have > seen, a complete failure. > > But hey, perhaps he can do it better. More power to > him if so. :-) > > May I suggest that you forward this thread back onto > list and respond there? > > - Sai >
I'll forward my response to your response, without snipping anything out of your response. Do you want me to forward my original message to you, as well? Tom H.C. in MI __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around


Jim Henry <jimhenry1973@...>