Why acronyms should not be predicates

[John Cowan]

I argue here that although acronyms (defined as two or more consecutive
letters) are currently interpreted as predicates in Loglan, it would be
better to interpret them as arguments.

1) Acronyms are inescapably polysemous

In Loglan, each predicate word is supposed to have a single well-defined
meaning. That meaning may be broad by the standards of any or all
natural languages, making the word vague, but it should not be ambiguous.
However, we cannot treat acronyms this way. If AmaBaiAma is defined
(in the dictionary, say) as "is the American Bar Association", what is
the predicate for "is the American Bookseller's Association"?

The only escape from this arbitrariness is to treat it as meaning "is
the ABA", whatever the ABA may be in context, which is precisely the kind
of polysemy Loglan is not supposed to have. Acronymfinder.com lists no
less than 98 meanings of "ABA", and says it has 250 more in the Acronym
Attic, a list of contributed but unverified acronyms.

On the other hand, single letterals are referential arguments, and
there is no difficulty at all with "Bai" meaning different things at
different times. The referent need not be explicit, but may be glorked
from context. Surely context is enough to tell us what CaiImaAma means
without dictionary definitions.

2) Acronyms lack natural place structures

Every predicate word must have a place structure. But there is no way
to deduce the correct place structure of an acronym predicate, nor any
real basis for doing so. In practice, most acronym predicates will mean
"is the <entity named by the acronym>". We already have a particularly
simple way of constructing such predicates: we can precede any argument
with "me". So if acronyms are arguments, and you want to say "is the
BBB", you can say "me BaiBaiBai" without loss of clarity.

It is true that we can convert a predicate of this form into an argument
equally easily with "le". But why say "le gaoFeogaoBeogaoKeo" when
"gaoFeogaoBeogaoKeo" is quite long enough? Indeed, "la Faibetakapas"
is shorter (in writing) and probably clearer.

3) It is arbitrary to have one rule for single-letter forms and another
for multi-letter forms

Enough said.

Note that I am not here considering the problem of distinguishing an
acronym such as NaiSaiAma from Nai, Sai, Ama which is three consecutive
arguments. That's a question for another day.

--
John Cowan http://www.ccil.org/~cowan co...@ccil.org
Your worships will perhaps be thinking that it is an easy thing
to blow up a dog? [Or] to write a book?
--Don Quixote, Introduction

[Randall Holmes]

I think the most important issue here is the one that John raises at the end, which I have already

addressed in my parser.

The use of letterals as pronouns is far more important than their use in acronyms, and so it is absurd

for us to have to pause explicitly between letterals in a chain of arguments.  Thus it is important that multi-letter acronyms *not* be pronouns as single letters are.  Moreover, the uses of acronyms are not as pronouns.

Further, I think the entire business of multiletter pronouns is a mistake -- it is very clear in NB3 in several places that JCB thinks that letterals are strictly single-letter, so there is no need to pause between successive letteral

arguments as long as we can tell them from an acronym.   I have solved this for acronymic predicates in my parser by requiring that an acronymic predicate be front marked with MIE.

I have a counterproposal to John:  I agree with his argument that acronyms should not be predicates.

I further suggest that what they really are are *names*, which means that they are usually automatically front marked with LA (and in fact I would propose that if acronyms were construed in this way that they should be required to be front marked with a name marker in all occurrences, so that they cannot be confused with strings of letterals).  An explicit pause would be required at the end of an acronym as at the end of any other name.  And this completely solves the problem of determining the beginnings and ends of acronyms.

This is *much* better than my current solution for acronymic predicates.   Moreover, one can still then say

Ti mela DaiNaiZa

One does not lose the possibility of using an acronym as a predicate.

I am so charmed with this that I think I am going to implement it.  My view is that multi-letter acronyms are simply a *pain*, and this is the least painful solution I have thought of.

--
You received this message because you are subscribed to the Google Groups "Keugru" group.
To unsubscribe from this group and stop receiving emails from it, send an email to keugru+un...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--
Sincerely, Randall Holmes

Any opinions expressed above are not the
official opinions of any person or institution.

[Randall Holmes]

An important point to state by itself:

John objects to single and multiletter forms being treated differently.

There is certainly some logic in this.

But I think it is necessary, for two reasons:

1.  strings of letterals used as arguments should not have to be broken with pauses.   Pronouns are much more important than acronyms, *whatever* acronyms are.   For this reason, we really need pronouns to be single-letter

and acronyms, if they are present at all, to be something different.

2.  The natural uses of letterals and acronyms are quite different.   In English, the mathematical x and the acronym DNA have quite different uses.

[John Cowan]

Randall Holmes scripsit:

> Further, I think the entire business of multiletter pronouns is a
> mistake -- it is very clear in NB3 in several places that JCB thinks
> that letterals are strictly single-letter, so there is no need to
> pause between successive letteral arguments as long as we can tell
> them from an acronym.

Well, I'm a programmer, and as such I am very fond of multi-letter
variable names (and consider the regression to single-letter type names
in ML and Haskell a defect). Also, multi-letter designating pronouns
can be very helpful when you are distinguishing between two people whose
names both begin with the same letter.

> I have a counterproposal to John: I agree with his argument that
> acronyms should not be predicates.

That was my main point.

> I further suggest that what they really are are *names*, which means
> that they are usually automatically front marked with LA (and in
> fact I would propose that if acronyms were construed in this way
> that they should be required to be front marked with a name marker
> in all occurrences, so that they cannot be confused with strings of
> letterals). An explicit pause would be required at the end of an
> acronym as at the end of any other name. And this completely solves
> the problem of determining the beginnings and ends of acronyms.

Well, that's okay, but then why bother with the grammar? "The CIA" just
becomes "la CaiImaAmas". I see no reason to add another grammatical
mechanism to save a single final consonant.

If that is set up as a convention, then Cai Ima Ama is three arguments
pure and simple.

--
John Cowan http://www.ccil.org/~cowan co...@ccil.org

There is no real going back. Though I may come to the Shire, it will
not seem the same; for I shall not be the same. I am wounded with
knife, sting, and tooth, and a long burden. Where shall I find rest?
--Frodo

[John Cowan]

Randall Holmes scripsit:

> 2. The natural uses of letterals and acronyms are quite different. In
> English, the mathematical x and the acronym DNA have quite different uses.

But the natural equivalent of mathematical x is rather "ba". In Loglan
the use of letterals is more like calling someone "Mr. S" when his name
is "Smith", and the multiple version corresponds to such name acronyms
as JFK, RLS, or GBS.

--
John Cowan http://www.ccil.org/~cowan co...@ccil.org

MEET US AT POINT ORANGE AT MIDNIGHT BRING YOUR DUCK OR PREPARE TO FACE WUGGUMS

[Randall Holmes]

The idea of simply eliminating acronyms (except for their use as dimensions) in favor of names terminated

with s occurred to me.   Its a quite radical and efficient solution!   There might of course be actual names from other sources that conflict.

It is actually not hard to implement the traditional acronyms as a special sort of name though -- it took me about half an hour to retune my parser to do most of it.

The issue of dimensioned numbers still remains, and is annoying for much the same reasons.  The solution

I installed in my current overhaul was as follows:

    I eliminated the front marker MUE that I imposed on dimensions, but I also required, just as for acronymic names, that one must pause after a dimensioned number  (this enabled me to have just one acronym class).  This removes problems with following letterals.

   There was another problem which is already known which was solved by the front marker and apparently reappears, which is the difficulty with "todai", which has two possible interpretations, one as "two dollars"

(or some other dimension D) and the other as "two of D" (where D is a letteral).  An unintentional reading

of the latter "to dai" as the former never occur if one pauses "to, dai" and can otherwise occur only if one pauses after "to dai" or if the "to dai" is final (the last seems most likely).

It is very clear in NB3 that JCB intended and relied on letterals being always a single letter.  I give a page reference for this in my Keugru agenda document.

On Fri, Aug 1, 2014 at 11:51 AM, Randall Holmes <m.randal...@gmail.com> wrote:

Actually, a lot of mathematical x's are pronouns, and not existentially quantified.   I speak math, and I think Loglan letterals are exactly analogous to a lot of uses of letters in mathematics;  the main purpose is to disambiguate reference of pronouns efficiently.

--
Sincerely, Randall Holmes

Any opinions expressed above are not the
official opinions of any person or institution.

[John Cowan]

Randall Holmes scripsit:

> The issue of dimensioned numbers still remains, and is annoying for much
> the same reasons.

The Lojban approach is to allow letterals in numbers, as long as they
are not initial; and by the same token, to allow digits in acronyms
(qua letteral strings) as long as they are not initial. So 2D is a valid
number, whereas D2 is a valid letteral-string. At least the first half
of this might be applicable to Loglan.

We use "boi" to terminate either a letteral-string or a number when
another letteral-string or number follows. This is inelegant, but it
doesn't come up that often (and is better than pause).

--
John Cowan http://www.ccil.org/~cowan co...@ccil.org

To say that Bilbo's breath was taken away is no description at all. There are
no words left to express his staggerment, since Men changed the language that
they learned of elves in the days when all the world was wonderful. --The Hobbit

[Randall Holmes]

I came up with a solution to the "todai" problem by reintroducing an optional  front marker for acronyms which has to be used in exactly the case of an acronym of length 1...an acronym has to contain at least two units, and one possibility for the first unit is the dummy MUE (the same as the marker I was using for dimensioned numbers).  So todai will always mean "two of D", as the only way to say

"two dollars" is "tomuedai".  One can always front mark an acronym with MUE but there is no reason to unless

it contains exactly one letter.  Two millimeters is <tomeimei>, no need for a front marker.

[Randall Holmes]

The cost of multiletter pronouns is a pregnant pause whenever two pronouns appear next to each other in an argument list.   Multiletter pronouns will be used *much* less than sequences of ordinary letterals, and are not important enough to justify disadvantaging speakers when the latter occur.

JCB had a different solution to this.   He noted that acronyms are penultimately stressed, and said that one

could disambiguate the strings of letterals by stressing them differently (finally, for example).   He wasnt talking about conflicts with multiletter pronouns, but with acronymic predicates, as he clearly thinks that pronouns are always single letters (in NB3) .  I think multiletter pronouns in TLI Loglan are an artifact of a program implementation error in LIP, actually.

[Randall Holmes]

Returning to the point about mathematical usage, look at any geometric argument:  the letters which appear are not quantified as a rule, but are simply being used as pronouns exactly as in Loglan.  The same is true in the texts of most mathematical arguments.  Of course, letters in mathematical expressions are often bound, and so would be ba analogues.   But it is more to the point to look at mathemetical English.

[Randall Holmes]

and JFK is not analogous in its usage to a letteral pronoun:  it is actually a proper name! 

[Randall Holmes]

and of course, like any name, may refer to a different person with those initials in a different context.  But this kind of contextual change of meaning of proper names is different from the change of meaning which is possible for letterals.

[Randall Holmes]

for multiple arguments with the same initial, we do have leicine, leicito...available.  We also have upper and lower case letters.

[John Cowan]

Randall Holmes scripsit:

> and JFK is not analogous in its usage to a letteral pronoun: it is
> actually a proper name!

Well, yes, in one usage. But it's also an acronym for John F. Kerry,
and an abbreviation for "just for kids", "just for kicks", and "just
from Korea". So the position I'm contending for is that when we
use "JFK", we are making reference to something else, which can only
be determined contextually. Names work for that too.

--
John Cowan http://www.ccil.org/~cowan co...@ccil.org

Go, and never darken my towels again!
--Rufus T. Firefly

[Randall Holmes]

Yes of course.   First names are an excellent example of the same thing:  la Tim is quite a different person in different conversations.

One can quite explicitly do something like

La Djan Keris ji la DaiKai...

and then use the initials DK in the rest of the conversation (or with less forethought, the second or third time the Secretary of State is mentioned, "la DaiKai ji la Keris")

I can see your argument for the occasional utility of multiletter pronouns.  But the phenomenon of explicit pauses between sutori letteral arguments is to my mind obnoxious.

[Randall Holmes]

So, I am definitely agreeing with you that acronyms should be arguments.  I think that they should be names, though, and clearly separated from the single letterals which are pronouns.  

There are *two* uses of acronyms, both of which present problems of potential ambiguity when they are adjacent to letterals (or numerals!).  The other is in dimensioned numbers.   I am proposing that both kinds should be terminated by explicit pauses; for the first kind this is natural, if they are construed as names.

The two varieties are both then naturally front marked:  I have set up acronymic names so that they are

*always* marked with a name marker in the draft parser which I will post later, and dimensions in dimensioned numbers occur...after numbers!  The todai problem (which already existed) is solved by providing an optional front marker for acronyms which *must* appear when an acronym contains only one letter.

I think that this is a much better solution than the one which I had in previous versions:  the basic insight that they ought to be arguments is valuable!

[Randall Holmes]

It will take a little while to post this as it require me to make changes in a number of documents and do some software tests.

[Randall Holmes]

I'm familiar with the Lojban solution.  I would include such an option to avoid pauses if the opinion of the Academy forced me to accept multiletter pronouns.  But it is awkward, in the same way that using pauses is awkward.

But I actually think that the situation with two or more letteral arguments after a predicate would occur moderately often.  It does occur in the corpus.

Another point is that it is quite clear what a single letteral pronoun is to be taken to refer to, and less clear how the reference of a multiletteral pronoun should be established.  I would guess that one might need explicit reference to what it is to be taken to mean, as in my John Kerry examples; and in this case one might as well use a name.

On Fri, Aug 1, 2014 at 1:15 PM, Randall Holmes <m.randal...@gmail.com> wrote:

I know about the boi solution.  I would actually include a terminator like that as well if I were forced to allow multiletter pronouns, as I prefer to avoid maneuvers which *require* use of pauses.

--
You received this message because you are subscribed to the Google Groups "Keugru" group.
To unsubscribe from this group and stop receiving emails from it, send an email to keugru+un...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--
Sincerely, Randall Holmes

Any opinions expressed above are not the
official opinions of any person or institution.

[Randall Holmes]

Notebook 3 is quite clear that letterals are single letters:

p. 54 last paragraph of 2.32   NB3

p. 101-102 NB3 is highly relevant.  The discussion here suggests to me what went wrong with LIP:  he says that the single letter elements of TAI are in one lexeme and the compound ones are in another --so there ought to be two separate classes, and there aren't!  (I have distinct classes).   He is very clear that multiletter forms

are not pronouns, but equally clearly the grammar in NB3 allows them to be read as such, because he does not actually create distinct single letter and multi letter lexemes.

p. 52 NB3 he explains how in his view stress can be used to avoid the need for pauses between letteral arguments.

He is quite right, as long as it is understood that letterals are single letters,  The point to my mind is that it is clearly his intention that pronouns are single letters.

[Randall Holmes]

I am *not* proposing to use the solution based on stress.  In fact, if acronyms are names there is no reason to assign any stress rules to them,

[Randall Holmes]

One final small point:  my optional front marker allows me to ensure that the space of compound TAI words

(as it were) can contain single letter acronyms without causing any ambiguity.  It is also usable with

acronymic names:  JCB's example of elemental sulfur is handled by la Muesai having the function of the chemical symbol S.  Since we are using acronymic names rather than predicates we say <Ti me la Muesai> for

"This is some sulfur" under my proposal.

[Randall Holmes]

I shall be quiet and in the next day or two finish fiddling with my PEG parser, amending some documents and doing some tests. 

This may impel me to write one of my Essays, the one on acronyms; if I do that I might come up with a formal proposal.

I would like to hear comments from others!

[Randall Holmes]

Formal proposals from others are also good, of course!

[Randall Holmes]

I make one more note, that though the *words* of NB3 say otherwise, the grammar given in NB3 and all subsequent Loglan grammars up to trial.85 *do* interpret acronyms as arguments where the context permits.  I believe, on the basis of the bits of text from NB3 shown, that this was actually a *mistake*, and not JCB's intention.

My parser does not accept multi-letter acronyms as arguments = pronouns, and I think this is true to the original intention (and removes the pause problem I have alluded to).