Friday, May 17, 2013

Uptalk

  I think I previously discussed my idea for an ASL/English/partial relex. I stumbled on some information for English's discourse intonation patterns since then and I have also thought about the general prosodic patterns, and I think I have an answer. Like many things, it's synergy
  English's uptalk pattern is interesting and somewhat novel and alien. I have thought about it before, but it seems ideal here after all, not just as a joke.
  The thing about uptalk, is how it is so full of filler and unusual rhetorical tricks. Vocal fry goes hand in hand. Someone said it sounds really aggressive and I would agree. It is ideal for a conlang because people revel in it, who use it, and it's empowering, and it holds the floor wonderfully. It's not some wimpy speech thing, it's demanding and pushy. When that's coupled with language learners demanding to be given the floor and hold it, it is both obnoxious and charming.
  The pacing is especially useful, because the rate of speech of someone talking is rather slow, often, and not that diversified. Using filler words like "thing(y)" and "stuff" is great for learners too. Uptalk and slang are where I would rather be, and I think they represent the turning of 2000 years of argument centering around explicit logical form and argumentation. Now it's back to informality and tribalism.
  Uptalk reminds me of questioning speech that invites the listener in, but also the continuation pattern in English, like in French at the end of prosodic words (coincidence?)--meaning it might seem like an invitation, but it's possibly just the speaker stalling for time, like a tag "um" or "y'know".

Wednesday, May 15, 2013

Intonational collocations, tropes & memes

  Hi all!
I'm trying to do more with the wiki again. It's just there's so much material. Do I start with Bolinger? Where do I start with the wiki? I wish I could get some sense of the overall structure. I think I will have to begin with ToBI's structure anyways. Arnold's notation is somewhat convenient but it isn't as interesting as it should be. Bolinger has 3 basic structures but lots of micro prosodic variation that doesn't show up again 'til we get back to IVIE or RAP e.g. pitch shifting.
  It's hard to get to a critical mass. I want to get enough going that other people can run with it and see it develop a life of its own...but other people won't do your work for you. I'm also concerned about making sure I don't borrow too much in fair use. Fortunately most of the ideas are easy to extend.
  What can other people do? A lot actually. Intonation should be intuitive; lots of people remember movie lines and can come up with more examples pretty easily when they hear one example. For instance listening to the 3rd column (HL%) here it is pretty easy to tell that the overall theme sounds very much like boredom or exasperation/irritation. It's not much stretch to hear a little kid saying some of those things, e.g. Ralphie from "A Christmas Story" or some of the sassy girls from "Mean Girls"; and from there to produce a specific quote someone can find via youtube or by renting a movie.
 Another concern is easily loading and hosting audio files from my device. SoundHost doesn't cut it because they want to play all the sounds in your album instead of just one at a time.
  I've found a very good book called "Prosody & Meaning" which has a fantastic article about lexicalization of pitch contours. As it turns out, they're essentially memes. This changes a lot of things, if micro prosodic variation is really so shifty. It means there's likely to be no underlying structure, and the rest is fashion. But, the authors point out that the same system that operates on focus and topicalization phenomena operates on the discourse functions, which include back channel responses, tags, etc. Which means that, very frequently, structures reappear just because without much meaning. They give the example of "meatball" and "school" showing up often with stereotyped patterns.
  They discuss other things like upstep and how it probably belongs in ToBI, and how pitch scaling may or may not matter. Other articles discuss how well-formed and discrete categories in intonation actually are. Turns out, not very. They're very slippery, and in fact TOBI categories appear to bleed into each other. Hence the usefulness of a exemplar-based theory of lexicalization, where lexemes come with a few available patterns. It also makes me wonder whether its worthwhile having a theory only so far as it helps organize the data. Too bad the data is so thorny, otherwise we might already have a TVtropes-styled wiki...Oh wait, I made one (it's just not useful yet.)
  Cheers!
 

Thursday, May 2, 2013

Intonation of English

  I should just change this blog's title to English Intonation.
  I've been reading things written in ToBI and it's hard. There just aren't great pedagogical resources for learning ToBI. The MIT course, this chart, etc.
  This article in the recently published 'Prosody & Meaning' is impressive in that it establishes the existence of a variety of lexico-accentual collocations (aka tropes) in the English vocabulary. This is significant; it means that intonational theory doesn't have to focus on abstract examples as much as exemplars, which tend to have very specific meaning tied to them. This is an approach to grammar I really do enjoy, and you see it pop up in conlangs sometimes e.g. Kalusa. They also point out that the accents, besides being prominent in terms of discourse functions, also recur in ordinary words; Calhoun, the author, points out that 'meatball' and 'government' have very particular and common contours that don't seem to have a discourse function; that's just how they're most often said. This is a fantastic insight, & it has all kinds of implications for learning languages. They didn't get too much into boundary tones (not at all in fact); they also pointed out various trouble spots with regards to ToBI analyses (e.g. speakers' ranges effecting perception of local maxima, upstep's neglect). The greatest boon was listing the tables of profiles & the words that fit into them. I think using the list could help make a writer more accessible, and I'm going to see if I can't get this kind of thing into TVTropes. These clever turns of speech are just too much to ignore; and though I think speech gets short shrift on TVT, I think this may be a chance to really improve that.
  I'm also looking into making a comic but I don't think I have the time or the will. I also don't have a story right now that isn't one of those "Gee whiz! Why don't I..."-derivative things. I think something is cool until I get down to it and I realize I don't feel strongly about it--or even writing, for that matter.
  I was looking into 3D software anyway and I realized there are a lot of mediocre webcomics that use CG. I think it's because they never spend more than 15 minutes on the visuals or the plot; perhaps too, CG is too demanding and they get roped into designing too much background. The characters are also unattractive.  Not too mention, emotionless. Unremarkable. The camera focus is usually in need of a lot tweaking. CG is one of those things where I feel like there should be several drafts and therefore several people involved telling you what looks bad, at least. It shouldn't be too hard to do those drafts on the other hand though. If there's one thing about CG, it's that no-one appreciates how much really work actually goes into it. No-work comics with stick-figure art usually at least have punch lines though, and most CG comics are serious rather than humorous. The thing about both bad stick art and mediocre CG is that they're both really consistently something on one level (for stick art, it's its inconsistency) while for CG it's consistent but it's consistently not what it's supposed to be representing.
  Hey, there isn't any reason why a very mixed style a la Klimt shouldn't work. Or Sealab.

Friday, April 26, 2013

Gregg Shorthand, learning

  I was preparing a version of this already, but I didn't like it. I only have time to do things piecemeal, so anyone reading this, I apologize for my posts from now on, they'll all be shorter and more piecemeal. It'll be like putting together a puzzle's pieces.
  I've been learning Gregg Shorthand for awhile, in a piecemeal way. I write fast but I don't know all of the theory, so I need to get faster.
  The basic way to learn the alphabet and the theory is to pretend Gregg is another alphabet, that comes with ligatures and abbreviations. Getting fast at translating eventually transfers to recognizing the shapes themselves but this is, I think, an important in-between step. It would be easier if people learned the abbreviations first actually.
  Cue this idea. There's a Gregg phrasal/entry dictionary that lists Gregg words in ASCII format. It's great but it's a little hard to read because it spaces all the graphs, digraphs, and trigraphs with dashes. If it just did it like the normal written language the abbreviations would be easier to learn; for example, the phrase "he can go" is just <ekg> (I know, ha.)
  ru ar oot ad haPND? 
  Are you <ru> aware <ar> of what <oot> had <ad> happened <haPND>? 
  Are you aware of what had happened?
  DTe g eththm OR n?
  Did he <DTe> go <g> with them <eththm> or <OR> not <n>?
  Did he go with them or not?
  This would be super-easy to use for a lot of texts if I could just get an automatic translation key working.

Thursday, April 25, 2013

What is a reddit?

When I said awhile ago that I needed a very special kind of news-site, I didn't realize that reddit fit the bill to a T. It's crazy, people create their own forums (aka subreddits) on the fly! All that's missing is a wordfilter thing like lernu.

Wednesday, April 24, 2013

Pirahã, Simplicity, & ASL

  Article about ditching recursion in clause subordination for "ontological promiscuity" aka storytelling.
  My head hurts. I've just basically learned that storytelling, rhetoric, and metaphor are all interchangeable tools, and I've been trying to learn language (specifically Gaeilge) with Where Are Your Keys? and I know I'm not getting it. It's just that when I look at the set-up I see all kinds of different things being brought to bear on what should be a simple situation but which isn't. Even the use of demonstratives i.e. 'this' and 'that', are not simple. Why don't we just use 'it'? What's the point of the game? The most difficult part is that it switches between role-playing, where you have to give a response, and copy-catting, where you're passively learning. I still don't get the hang of transitions. It's probably because I've never really actually played with another good player. It's just not viral. What's the point? I'm talking about coffee? And then it hits you, "Yeah, I'm talking about coffee in another language and by playing along other people can too!" That is the craziest feeling ever, but it entails a loss of will you don't really feel unless you're already an author, I think. It's very hard for poets to do that, because it means sinking into a language rather than carving it out like poets do, creating an antilanguage. You don't have to understand exactly what's going on, or that there are several options available to you (I think those come up later, in more fluent speech, and aren't too tricky? But maybe I'm wrong. I found variations in the use of this/that based on empathy and perspective in English; crazy but cool! And they're just a casual everyday part of the language!)
  This is why I wanted to dig around for a language that truly needs no more distinctions than just pointing, naming, and describing. I'm talking about ASL. Why not create a language that does for ASL what ASL did for English when they made Exact Signed English? Basically, bend English to ASL's rules like when they bent ASL to English's!
  ASL has no words for articles, like Latin. Unlike Latin, and like Chinese, it has classifiers, and a whole host of other exotic phenomena. Notably, it tends to drop the inanimate pronoun (relying on context) and of course, deixis and pronouns are merged under ostension (aka pointing at stuff) so there's only pointing at yourself or other people. English is a German language, which has the least dropping of any language family, but that is changing somewhat in America (where pronouns tend to be left off the beginning of sentences sometimes); other languages like Japanese and Piraha tend to not need pronouns at all except as almost like nonce titles (like 'I' actually meaning 'yours truly', etc.) On the other hand, Japanese has a very complicated animacy hierarchy which does a lot of work circumscribing what cannot go into a sentence's interpretation. These languages which do not need pronouns are in the minority.
  The ASL situation actually reminds me of the deixis in some languages, which marks for near, far and out of sight; similarly, we have me, you or him or her, and the people I can't gesture at. The latter are particularly interesting, because they are the language's other innovation: signing space. Basically, this little piggy goes to market, and this little piggy...The storyteller points to spaces he has established as being occupied by imaginary entities, of which there can be at most 8.
  This trick takes advantage of natural tendencies of the human mind, e.g. the Roman memory trick. It also breaks down very difficult recursive sentence structures and turns them into stories. Pirahã does it super-effectively to get around using clauses in any sort of difficult fashion. It seems like the stories unfold in a way that makes it easy to keep track of the referents, but not that easy. In fact, this shows why clauses exist at all, to make telling the order of events in the way you want, easier. On the other hand, subordinating parts of sentences means that some phrases and clauses are not going to have the same bearing as other sentences, they're going to serve other sentences. This has an effect on the breaks, or cesura, between clauses. To the Pirahã, this is perhaps unacceptable. The Pirahã don't just speak their language like other peoples—they mumble it, they whistle it, etc. There's even an effect whereby the few consonants in the language can swap out for each other at different times. You cannot make this up; its called the 'Sloppy Phoneme Effect', whereby what matters to the speakers is the syllable structure in terms of tone (Pirahã is a  tone language like Chinese), stress, length of vowels and length of consonants. The language is highly musical as Keren Everett says.
  What this means, and I'm just guessing here, is that Pirahã prosody is a pain. It's fully loaded with all the distinctions the language makes already, so it cannot carry the additional burden of, like English, telling speakers what sentence focus and informations structure is—so they resort to a highly conscripted definition of a clause and a sentence, and don't differentiate between different levels. Perhaps it's not so much that they cannot speak, but that all the prosody that can be carried over into whistling, etc. determines the baseline speech. In other words, the whistled speech could be the bottleneck for the spoken speech—which is unprecedented. It's like if texting determined speech, and everybody stopped talking on their phones and just—oh wait. But in truth, the written word has functioned over the years...much the same way. How much gets left out from the written word, when compared to speech? Yet we use it, in the past to write letters. Our words carried over long distances to reach others' ears. How is a whistle any different?
  This post is still about ASL, the only language that is still more easy to understand orally/visually than in any form of writing. It is also, perhaps, the one language I still really care about, because you can talk about things you see and experience. It seems like the purview of all language nowadays is writing, and fiction at that.
  I thought it would also be cool if a sign language became oral like is suggested as the source of all languages by some researchers. That could mean that a conlang could really take off the ground, and that there may be many origins of language!

Intonation; ToBI Revisited; Where are your keys?

  I believe I've spoken before on the discontents of using ToBI before and briefly mentioned some of the alternatives (the British school, Bolinger, Rhythm & Pitch or RaP, etc.)
  I personally think it's challenge lies partly in its presentation; if you could condense it into punctuation and diacritics people might like it more and it would be easier to use (not to forget, punctuation in English seems to focus mainly on the "cesura" or breaks, much as the Break Index part of ToBI does, but people have been using it forever, and its not conventionalized but it serves its purpose well enough for the times and the fashions…but I digress. Another time).
  ToBI is tricky too because of the lack of audio recordings and quality presentation. So can you imagine my surprise when I found this website? It's absolutely fantastic because for the first time ToBI is presented succinctly and with a clear sense. I can finally hear what's behind it, and its not so bad after all. It seems fairly well-organized albeit focusing on rather pronounced and unusual distinctions, but I like it. I also notice that there are several correlations with the British tradition, in that there are low falling, low rising, high rising, etc. patterns, though in ToBI they represent phrasal interactions. One remarkable thing is that the British "Mid" seems to match the "High flat" and that there isn't a corresponding category to the British "High fall" (reference: John Maidment's NEWTON program, online; though it must be noted that the British tradition doesn't have a distinction between, for the most part, between the boundary tones that I just mentioned, and pitch accents.)
  On the other hand, there are pitch accents which more nearly correspond to the idea of "nuclear tones" in British. There is one final stressed syllable in every intonational phrase (but what is an intonational phrase? other researchers have suggested that we focus on breaks, continuation patterns, and cesura to better understand that, but again that's for another time); this bears the stress that truly matters. That's what the British tradition says, and that may well be true—for the British. However, the appeal of ToBI lies in being able to distinguish between say, for example, a "peaking"pattern that can spread left, and one that can spread right. The stress falls on a specific syllable, and the syllable aligns on its left or right side—this is remarkable, and compelling moreover. (Certain words, like 'America', generally have boundary tones on both sides of the pitch accent; they're special, but ToBI can still work for them.)
  It surprises me that there are only 22 combination tones (or is the page mistaken? I am not an expert anyway). It's almost like we can just have 22 punctuation symbols. Awhile ago, a Frenchman tried to introduce extra punctuation, for instance a symbol (in place of a period) for endearment, another for sarcasm, and so on. (Entertainingly the romantic pronouncement mark looked like a little heart.) I don't think it's a bad idea at all, though it would mean people would have to do one more thing to read well—maybe it would catch on, if people could be educated as to its systematic use. It would be brilliant for texting.
  The problem with intonation, though, is that there seems to be no end. It's apparently bottomless. The distinctions made in ToBI in one paper were found to be gradient rather than categorical, meaning that they bleed into one another with wide overlap. They compared them to the vowelspace where they all bleed into one another, and that may be true, but…I feel there is a lot more. The quality of the peaks and valleys, their acceleration , gradation, etc. do not fall under the purview of ToBI; I think a resynthesis and integration of the kind of insights I saw in Bolinger, with the appeal of ToBI, would work wonders.
  Whenever I read the British schools' works I can't help but read it in that dramatic, over-accented Downton Abbey-esque bombast that the actors over the pond seem so fond of.
  I read that there were attempts at integrating the British and ToBI but it cannot have been too successful, they split hairs differently. I'll still have to go look for them again though.
P.S. I should be able to potentially use ToBI to pick out tunes to use in role playing. This should be especially useful for making plausible scenarios happen during Where Are Your Keys? gameplay, which currently vary amongst "game masters" as to the communicative goal of various interactions. These vary from language to language and are tricky, but I think just focusing on them will bring the right focus to gameplay, and what it should be about—in other words, functionl, pragmatic use of intonation to make small distinctions and split hairs in the ways natural languages do (split hairs, that is.)
P.P.S. I can finally get started on my Intonation Wiki that I've wanted to work on for so long! The ToBI examples are great, and I can easily think of scenes from movies that they pop up in.