Demonstrative deixis in Catalan

Once upon a time (a time which still continues further south towards València) Catalan had a three-way deictic opposition in its demonstratives analogous with este/ese/aquel in Spanish and hic/iste/ille in Latin: that is, with one form corresponding to each of 1st, 2nd and 3rd person. This scheme is outlined below:

pronoun adjective adverb
1 açò aquest aquí
2 això aqueix allí
3 allò aquell allà

So in this system, aquí = "here by me," allí = "there by you," and allà = "over there by someone else." This obviously differs from the English system in that 2nd and 3rd person have merged with us, giving us only here = "here by me" and there = "there by you" or "over there by someone else," which seems only natural, given that I'm a native English speaker.

Modern standard Catalan also merged two of its persons, though in a totally different way which confounds my English-speaking brain: 1st and 2nd person now share a form. This gives the following paradigm:

pronoun adjective adverb
1/2 això aquest aquí
3 allò aquell allà

This means that when you point at something your interlocutor is holding and ask after its identity, you have to say Què és això? = "What's this?" It's not too terribly difficult to imagine how this might happen -- as Amelia points out, until very recently any referents of això were likely to be in the same room, or at least within view, as opposed to referents of allò which would universally have been further away and possibly not visible. It's a reasonable, fairly natural grouping, on the face of it.

Things get weird in the modern age, though, particularly where the telephone is concerned. For example, since 1st- and 2nd-person deictics share the same form, you have to use aquí = "here" to refer to the location of the person on the other end of the phone! A question like Què passa aquí? is ambiguous -- it could mean either "What's going on here where I am?" or "What's going on over there where you are?" It's easy to imagine some ridiculous ensuing conversations: No, no, no pas aquí, sinó aquí!

So how does the ambiguity get resolved? You can't use allà because that would imply a place unconnected to either party in the conversation. I have to imagine that there must be some conventionalized circumlocution -- aquí amb mi or aquí a prop de mi vs. aquí amb tu or aquí a prop de tu or whatever. And if so, this would be incredibly ripe for grammaticalization, perhaps even yielding another three-way deictic opposition once the forms have fossilized and reduced phonetically and all that good stuff.

Of course, this is my English-speaking sensibility talking, which can't imagine not being able to distinguish lexically between locations on either side of a phone call. It would be interesting to observe how this really works in Catalan-speaking areas -- maybe they don't have a problem with the ambiguity. Good for my armchair-phililogist complacency to remind myself that "natural" and "reasonable" are extremely subjective concepts...

Koa online at last

Koa, an IAL project I've been working on for about nine years, has now got a blog devoted to it here. There's also a link to it over on the right, but maybe you hadn't noticed yet...

Anyway, this new forum is incredibly exciting for me, and I've been being unusually productive now that I don't have to hold all my new ideas in my head simultaneously while I think about them. Serious kudos to Amelia for the idea.

Loa la po sisi! ...keka ta halu ko opi ti?

(or would that be Keka ta lu opi ti? Something for the blog.)

NEVs and San Leandro living

In the unlikely event that there's anyone I haven't told yet, Amelia and I are going to be moving in together this summer. Or, more accurately, I'm going to be moving into her house in San Leandro. Yes, yes, I know, San Leandro isn't usually the first place on one's list of, say, vacation destinations, but the thing is that it's a really neat older house that she actually owns, with a huge back yard with fruit trees and oodles of garden space, and we're going to have bees and chickens and it's all going to be pretty spectacular.

However, San Leandro is still very much a suburb, and the nearest cute tea shop is in Alameda about eight miles away -- kind of too far even to bike on a daily basis, at least with my schedule. Grocery stores are also not exactly what I'd call "walkable," and though I don't personally shrink from the two-mile walk to and from the BART station, I imagine my friends would be more likely to visit me if said visit were not punctuated by said walk.

SO, if I'm going to be living in San Leandro, which I am, the conclusion is that I need a car. But I really don't want a car -- that is to say, one of those things that you pour expensive flammable substances into, the which you subsequently set fire to and dump the resultant smoke into the air in order to make it go -- so Amelia suggested that I look into Neighborhood Electric Vehicles. I'm not totally sure if she was serious, but the more research I do, the more it seems like this is exactly what I need. For longer (or faster) trips we can figure out how to trade her Camry (or maybe Prius someday soon, please please please?) back and forth, but for the vast majority of driving I would be doing, a cap of 35 mph and 40-45 miles would be absolutely fine...especially considering the fact that it would be virtually free to operate.

After a bunch of comparison shopping, it looks like the Kurrent by American Electric is my best bet. And it's nice that I would actually be able to afford it with my new job.

I don't know for sure if this is really going to happen, but I'm pretty excited about the idea. Some photos:

Gah!

Trying to record oneself playing a musical instrument is officially the most frustrating activity of any kind ever. I swear I can play Valse à Bruno with my hands tied behind my back ordinarily, but with the camera rolling I can barely get through three measures without making the kinds of mistakes that are just too bad to soldier through. After half an hour of trying, Ossie and I are spent. Arr, gah, bleah, eeeeeeee, and other noises of exasperation.

The Celtic Bandoneon

I'm trying something new here. I really want to dig in and figure out my bandoneon so I can start doing all kinds of cool complicated two-handed stuff, but I think I need to slow down and concentrate on getting a feel for the arrangement of the notes. For the moment, then, I'm going to try focusing on one hand at a time, playing all the Irish tunes I know -- they're fun, short (for the most part) and diatonic, which is the perfect combination.

So this is my first attempt to play "Scatter the Mud." It's in something like Eb Dorian because I feel like the usual A is a bit screechy on the free reeds. I'm going to try really hard here not to be apologetic (except for forgetting the repetitions on the B part -- oops) and just say that you can basically tell what I'm trying to do, and I look forward to posting another video when I've really got it figured out. Anyone interested in backing me up with guitar/bouzouki/fiddle/tin whistle/Irish flute/bodhran/kazoo?

An input-optimization puzzle

As I was walking home the other day I was thinking about the way predictive text input works on cell phones, and how it's simultaneously cool that, because of the fact that only a tiny frequency of the possible combinations of our 26 letters are actually used as words, with only 8 symbols we're capable of entering English text without a great deal of difficulty, and irritating that a lot of common words have duplicate input sequences: soon/room, of/me, he/if, etc.; and the less frequent but still aggravatingly abundant book/cool, mind/mine/nine, cast/cart/bart, good/home/gone/hood/hoof/hone/goof, and on and on.

For reference, the way my cell phone (I assume this is a standard at this point?) groups the letters is as follows:

2 - abc
3 - def
4 - ghi
5 - jkl
6 - mno
7 - pqrs
8 - tuv
9 - wxyz

Well, I found myself thinking, what if we weren't required to disperse the letters across the keys in alphabetical order? Couldn't there be some kind of optimized arrangment such that the numbers of duplicated sequences would be dramatically reduced? I figured that I would be in a good position to try this out, having spent a lot of time thinking about the frequency of letters in English.

To this end, I read the TWL (US Scrabble) dictionary into a database and set up some automatic calculations to derive cell-phone input sequences for all the words. As it turns out, the standard layout has about 27,000 duplications out of the approximately 163,000 words in the dictionary (16.5% duplication). This shouldn't be hard to beat, right?

As it turns out, my optimism (and ego) was unwarranted. My first attempt gave me 33,000 duplications (20%), significantly worse than the baseline. After about a half hour of tinkering I managed to get it down to 25,000 (15.3%), which salved my wounded self-confidence but still was not nearly as good as I had hoped.

There must be some principles that are going to underlie any successful attempt. One is clearly that no two vowels (including y) should appear on the same key. Another one that I embarrassingly only discovered after the aforementioned half hour is that consonants frequently used in derivation and inflection should also not coincide, notably d, s, r, n.

So here's the challenge: can you come up with a scheme that's better than the 16.5% baseline standard, and/or better than my 15.3% best so far? If you send me your letter groupings, I'll feed them into my program and post the results here. Maybe together we can figure out the Grand Unified Theory of Duplication Avoidance in Cell-Phone Text Entry. Or something.

Pass the meme

Given that I've now seen this both on the Bluff post office wall and on Oona's blog, I can officially class it as being EVERYWHERE, so I'd better post it too so I'm not left out.


I would just like to say that I would live here in a heartbeat. Does anyone know the story of this...uh...complex?