Near-optimal Loglan syntax

From Lojban
Jump to navigation Jump to search

Jeff Prothero's article "Design and Implementation of a Near-optimal Loglan Syntax" in which a language called 'Plan B' is described (May 1990). The article prompted a satirical response in Jacques Guy's 'Plan C' (Sept. 1992). If any language results from considering Plan B and Plan C, then I guess it ought to be Plan D. Also arising from the first of these discussions Jörg Rhiemeier developed his X-1 language.

What does Plan B do & not do?

In fact, 'Plan B' does not give a syntax for what we now call a loglan (a language based on the principles set out by James Cooke Brown) or even a loglang, in that it does not implement formal logic. Indeed, it seems clear from what Jeff Prothero himself wrote that Plan B is what we now call an engelang; cf.: "Here I will propose a loglan syntax which:

  • Is simple enough to be parsed by a couple of hundred lines of straightforward C. (See attached program.)
  • Is simple for humans to learn and use.
  • Allows for unambiguous resolution of continuous human speech.
  • Offers near-optimal conciseness and simplicity."

'Plan B' certainly fulfills the first criterion. This, however, has no direct relevance to a loglan(g), but it may, of course, be the aim of an engelang. However, I am (and I write as a retired computer scientist, whose specialism was programming) firmly of the opinion that computers are tools to make life easier for humans, not machines for which we should make 'life' easy.

As regards the second criterion, there are, in my opinion, features about 'Plan B' which do not make it easy for human use, e.g. the bizarre spelling system in which each letter may be pronounced as a vowel or as a consonant, that every morpheme has two quite phonetically disparate allomorphs, that morpheme segregation is effected by a "Huffman-style expanding-opcode sort of scheme". It is my opinion that the first criterion has been achieved at the expense of this second one.

From the examples which Jeff Prothero gives in his article, it would seem that 'Plan B' does offer an unambiguous resolution of its input. But, as far as I can see, this is achieved by relexifying English and providing it with a syntax so that it can be unambiguously parsed using a binary tree. Whether this makes for "near-optimal conciseness and simplicity" is, I think, a matter of debate.

In short, what Plan B does, it seems to me, is to implement a method in which:

a stream of characters (whether alphabetic or of binary digits or, indeed, any other characters such as Tengwar) can be unambiguously broken up into morphemes or 'affixes', as the author calls them; the 'affixes' can be reconstituted as a binary tree.

Thus, Plan B provides a way whereby one may analyze an English sentence as a binary tree and then generate a continuous stream of characters (alphabetic, bits or whatever) which both maintains the same word order as English and unambiguously represents that tree. I leave it to the reader to decide if the language is justly parodied by Jacques Guy's 'Plan C' or not.

Why I began and have abandoned the project

The discussion in September 2005 concerned itself mainly with the orthography and phonology of 'Plan B'. In March 2006 I returned to this and developed a radically different solution: a syllabary of 16 symbols. I also considered the question of 'self-segregating morphemes' and the 'Plan B' grammar. In doing this, I had originally thought to produce a loglang of my own on the basis of these considerations. But it became clear to me that any loglang syntax would have to be radically different from Plan B's syntax. With, however, its restriction to 16 symbols, the results were likely to feel very kludgey, especially when compared with existing loglangs such as Loglan, Lojban, Gua\spi, and Ceqli.

In March 2009, during threads on the Conlang list related to oligosynthetic languages (the original thread had split into two or three related ones) , I wrote:

I recall that in the late 18th century Father Joachim Bouvet sent Leibniz a letter in which he described the I Ching (yi4 jing1). He listed the hexagrams not in the weng-wang order in which they are commonly listed, but in the Fu-xi order. Leibniz read the hexagrams as the numbers 0 to 63 in binary notation! So why not 2^6 concepts? 64 'primitives' would still constitute a challenging oligosynthetic language.

From this, I conceived the idea of expanding the syllabary of 16 syllables to one of 64 syllables, each being written with a hexagram, and thus to develop an oligosynthetic language. In the end, however, I gave this up because:

I think 64 primitives are too few in which to express the whole range of huuman thought, experience and knowledge; a quick look on Google will reveal that the Yì Jīng (I Ching) hexagrams have many cultic associations and, it seems to me, that no matter how much I may dissociate the language from these, mystic and/or cultic associations will remain; also an oligosynthetic language, even if possible (and I have grave doubts on that score), is a long way from a loglang or loglan and would have nothing to do with Plan B (or Plan C).

Indeed, I have become more and more convinced that any reform of Plan B would make the whole thing collapse; thus, as the old nursery rhyme says, after B and C, D has tumbled down. And the cat in the cupboard? That, methinks, is the one Jacques Guy let out of the bag with C.

However, I leave these pages online in case anyone is interested.

Discussion

Another loglang, with design goals similar to those of lojban but done poorly. Elegantly mocked on the Conlang list, reproduced below. The serious information about it is stored here


  • .kreig.daniyl.:
    Ironically (since I am the one who put this satire of it on the wiki) I actually found Lojban indirectly through Plan B. It seemed an insane language, so I wondered whether the existing loglans it refered to were any good. So I googled Loglan and found mentions of both Lojban and TLI Loglan - including a website comparing the two. Lojban, like Plan B, had all the features that had attracted me to Esperanto until I found out that Esperanto didn't have them.

Searching for a snail-mail address I had mislaid, I had to revive my

mailbox and wade through 993 messages. As I did, I chanced upon

several which I must have missed. One was telling us about the

availability of a certain "Plan B" at hebrew.cc.columbia.edu. 'Twas

like waving a red rag at a frog (a bullfrog) and I snapped at the bait

merrily.

I read Plan B on the taxi ride to the city on Telecom business which,

according to Telecom, you have no business knowing about. But Plan B

is no corporate secret, so I want to tell about Plan B "Design and

Implementation of a Near-Optimal Loglan Syntax". Be warned: I'm about

to take the mickey out of Plan B from Hebrew Space. And why shouldn't

I? It reads as if the author was having us on. So it's only a mickey

for a minnie see footnote 1. Nevertheless, Plan B gave me a great

idea for the perfect conlang. Just bear with me as I unveil before

your amazed eyes the arcane mysteries of Plan B, you will be fully

rewarded in the end.

The Plan-B language -- I'll call it Bee for short -- Bee, then, has 16

er... phonemes, because sixteen is a power of two, which makes it

computationally desirable. Each phoneme has two allophones, one of

which is a vowel, or a diphthong, or the same preceded by "r", the

other a consonant. I say: jolly good idea! Indeed, it's like the

author says: "By providing both a vowel and a consonant pronunciation

for each letter, and using them alternately, we can pronounce

arbitrary strings of letters without difficulty". Brilliant. And I,

poor sod, who thought a strict CV(V) language would do it!

For instance:

English: "I like her driving my car"

Bee: "G-l tk-s ck-l mg-n g-n cc-l"

IPA: g-rE ti:-s eik-rE mai-n ai-n eiS-rE (S = esh, E = epsilon)

This is terrific, for it means that each word has exactly two

allomorphs, depending on how many er... phonemes precede it in

the sentence. "Ck" for instance, is either eik or Si,

e.g.

English: "She likes me"

Bee: "Ck-l tk-n g-l"

IPA: Si-l ruk-ri g-rE

The unambiguous segmentation of the spoken chain into its

discrete words, implemented oh! so very messily in Lojban,

is implemented Huffman-style in Bee. And I, poor sod, who

thought the first vowel of a word could be used to tell

how many syllables it had!

Let us now turn to the grammar of the language. It makes do

with an unlimited number of ... er... case-markers, of which

you have already encountered three: -l, -n, and -s. -l has

highest precedence, -n second highest, -s third. Armed with

the vorpal sword of that knowledge, you should be able

to disentangle the Gordian knot of the two sentences

above in even less time flat than Alexander.

No? Well, perhaps two more sentences would help:

I drive the car

G-l mg-n hb-n cc-l

I can drive a car

G-l cn-n mg-n b-n cc-l

Ha, ha! I hear you say, why "Gl cnn mgn bn ccl" if "Gl tks

ckl mgn ccl"? Shouldn't it rather be "Gl cns mgn bn ccl"?

I agree with you. It's probably a typing mistake: "s" and

"n" are rather close together on a keyboard. However:

I will drive my car to you

G-l ml-n mg-n g-n cc-l th-n j-l

So, clearly, it wasn't a typing mistake. Now where have I

put that vorpal sword again?

In conclusion, the author sums up the advantages of such a

language. I cannot resist the pleasure of quoting ckl:

Compared to existing loglans, Bee

  • Is much simpler.
  • Potentially allows for mechanical recognition of continuous speech.
  • Is suited to laboratory studies of the Sapir-Whorf Hypothesis.
  • Possesses a certain elegance. (Eat your hearts out, Coco Chanel, Christian Dior, Pierre Balmain and sundry!)

Well, folks, allow me to present my own loglan, beyond Bee.

So I'll call it Cee.

Cee is written in an alphabet of 26 letters: a, b, c, d, e, f... I'll

leave you to guess at the rest. Those letters are pronounced

respectively bi, ba, sha, da, fi, fa... I'll leave you guess at the

rest. That leaves us quite a few handfuls of syllables out of which we

select:

bo as first-order precedence whatever, which we write 1

sho as second-order precedence ditto, which we write 2

do as third-order of the ilk, which we write 3

... I'll leave you to guess at the rest.

And also:

bu which we write <space> or <spaces>

shu which we write .

Now look:

English: I drive the car.

Bee: G-l mg-n hb-n cc-l

Cee: Me-1 drive-2 the-2 car-1.

(Hyphens have been inserted only for your convenience, o, gentle readers!)

English: I can drive a car.

Bee: G-l cn-n mg-n b-n cc-l

Cee: Me-1 can-2 drive-2 a-2 car-1.

I'll leave it to you to work out the pronunciation of those two Cee

sentences. Just note how that little syllable, bu (spelt <space>),

neatly and *elegantly* solves the problem of recognizing morpheme

boundaries. Cee, admittedly, is more verbose than Bee, but I'm working

on it. Now where is my copy of Dutton's Speedwords?

Footnote(s)

Note 1 Why "minnie"? Well, we all know who Mickey and Minnie

Mouse are, don't we? And what sex they are. Yes? So, it's...

twit for twat!

Links