Message-Id: <9207010430.AA09723@relay1.UU.NET>
Date:         Tue, 30 Jun 1992 23:03:17 -0400
Reply-To: Logical Language Group <cbmvax!uunet!GREBYN.COM!lojbab>
Sender: Lojban list <cbmvax!uunet!CUVMB.BITNET!pucc.Princeton.EDU!LOJBAN>
From: Logical Language Group <cbmvax!uunet!GREBYN.COM!lojbab>
Subject:      fyky, young men, televisions, negatives in attitudinals
To: John Cowan <cowan@SNARK.THYRSUS.COM>,
        Eric Raymond <eric@SNARK.THYRSUS.COM>,
        Eric Tiedemann <est@SNARK.THYRSUS.COM>
Status: RO

I haven't carefully read the comments that have been posted, but soomething
caught my eye as being worthy of immediate response rather than waiting
for me to catch up from now being 2 months behind in mail processing...
From the phone game discussions, with Mark, Nick, Ivan, and Colin
expressing various positions (From several messages, so don't match up
number of indents with single individuals):

>>Mark's starting sentence was:
>
>>Until you start sitting up straight and stop playing with your food, young
>>man, there'll be no television for you --- that's for sure!
>
>>He handled this as:
>
>>pu'o le mu'e do co'a xagysirji zutse gi'e na'e kelci le do sanmi doi citno
>>nanmu kei .i'enairo'a do cu .e'anai catlu le se tivni vau ju'o

>Yeah, that {e'anai} caused many problems in the TV sentence.  I'll concede
>that what I wound up saying was probably not the right way to convey the
>meaning, but it's interesting to look at why I did.
>
>I had {do cu .e'anai catlu le se tivni} for "There'll be no TV for you."
>
>> >  Does it mean "(I forbid this!) You are watching TV",
>> >  or "(I forbid that) you should watch TV"?
>>
>> The first.
>>
>> >  {mi cmima .au le fonxa kelci} - "I want to be a member", or "I am a
>> >  member, and have wanted this"?
>>
>> The second.
>>
>> No interjection can change the meaning of the sentence _I am a member_
>> to `I am not a member' or `I want to be a member'.  And attitudinals
>> are just that, interjections.  They show your evaluation of the fact
>> reported, but don't alter the fact that the fact is reported.
>
>I agree entirely. [I think I have not followed this in the past, but I
>endorse Ivan's opinion and intend to follow the precept henceforward.
>"ko na tavla .e'anai", not "ko .e'anai tavla", for "don't speak".]
>
>> In general, my strategy as a fonxa kelci is based on the understanding
>> that the sentence(s) I get from Mark are to be translated faithfully,
>> even if real world knowledge suggests that he intends something else
>> than the obtained meaning, and that his additional comments must not
>> be read until my own translation has gone off to Colin, or if read
>> must be ignored and not allowed to influence my understanding of the text.
>
>I too have followed this strategy (though I confess I have sometimes
>allowed guesses about the English to govern the style of my translation,
>but not the content).

I have to disagree strongly with the apparently evolving consensus.

An 'interjection' CAN significantly change the meaning of a sentence in
Lojban (as well as in other languages).  Only a very small subset can
make a statement mean its exact opposite.  In English, of course, there
is the recent phenomenon of postposition "NOT!"  Yiddish is full of
colorful interjections that alter the meaning of the sentence, often
into its exact opposite - I remember a particular discussion very
reminiscent of how selma'o UI works in the beginning of _The Joys of
Yiddish_ showing how where an attitudinal appears in a sentence can
strongly effect how it is interpreted.  Same with Lojban, which took a
few cues from Yiddish.

"ko na tavla e'anai" cannot mean "Don't speak".  How would you express
using attitudes "I forbid you to NOT speak".

Mark's original was probably correct or close to it.

A hopefully brief discussion:  All of the members of UI are in the
category of metalinguistic comments.  They are expressing ideas, claims,
or in the case of attitudinals, emotional reactions to the relationship
expressed by the sentence.  It is not necessarily the case that a
relationship be realized for one to need to be able to talk about it
metalinguistically.  To insist that any statement of a bridi must
logically claim it to be true means that you cannot talk about
falsehoods metalinguistically.

The obvious example of a metalinguistic that does not claim a truth is
"xu", the yes/no question.  "xu do klama" (Are you going somewhere?)
does not claim that the answer is 'yes', or "go'i".

Most of the attitudinals and discursives indeed have no effect on the
truth value of the sentence.  The ones that do, generally do not
directly contradict the sentence, but rather comment on the hypothetical
world wherein the sentence is true.  In other words, they form a kind of
subjunctive.  THe people in this discussion have already discovered that
"da'i" ("supposing") is such a UI.  "da'i" does not claim that the
attached bridi is true - and indeed, by contrast with "da'inai" ("in
fact") may seem to imply that it likely is NOT true.  But in any case it
indicates that the speaker is talking from the point of view wherein it
WERE true.

".ai" (I intend) is also an obvious 'subjunctive'.  The relationship is
probably not true at the time the speaker expresses it, but the speaker
intends that it come to be true.  ".au" and ".a'o" (I want and I hope,
respectively) are rather weaker - they effectively eliminate the
truth-functional nature of the underlying bridi - it doesn't matter
whether they are true or not - the speaker is expressing reaction to the
hypothetical 'if it were true'.  (I note in passing that the
hypothetical "if", while not the same as the truth-functional
"if...then", works quite similarly:  if it is false, then any follow-on
claims are moot and do not affect the truth of the combination.  If the
antecedent is true, though, then the consequent must also be true for
the whole to be true.)

".e'u" suggestion, and "e'o" petition also make hypothetical worlds.
When we say ".e'osai ko sarji la lojban." we are not making claims that
all readers DO or even WILL support Lojban - we are petitioning them to
make it true.

Indeed, most of the members of UI in the .aV, and .eV series move one
into the hypothetical world to some greater or lesser degree.  The .iV
series were the toughest ones - ".ianai" has been used to express
disbelief of a truth as well as ironical counterfactual.  .oV and .uV
attitudinals seldom affect whether a claim is being made.

Indeed, when I shifted the words around in redeveloping the set of
attitudinal set, this was one major criteria.  For a while (1987-88)
when we first developed the attitudinal system, the initial letter was
to be an indicator of a type of attitudinal in this respect.  We found,
however, that there wasn't an equal number of attitudes in each of the 5
categories of attitudes that we came up with - the .uV and .oV words
filled up rapidly, and started encroaching on the .iV space.  We also
found that each of the "'mental world' attitudes' in some cases had to
be interpreted as "real world" attitudes.

By the time of the draft textbook, the paradigm of categories was badly
strained, and I did not teach it to the first class.  Indeed, the 'real
usages' that started coming about during and after that first class
planted the seeds for the final revision of the attitudinal system that
is described in the attitudinal paper.  For about a year, there was a
cmavo that explicitly declared whether the attitude was 'real world' or
'mental/hypothetical world'.  THis word was never used - and may have
been the convincing proof to me that attitudinals had to rely on
simplistic expression, rather than calculated analysis.  It disappeared
with the attitudinal paper, though it lives on as a ghost in ba'u
(exaggeration/understatement) je'u (truth/falsity), ju'o
(certainty/uncertainty), and evidential ru'a (postulation - which
unambiguously moves one into the 'mental world').

Summarizing:  the insertion of attitudinals may amend or eliminate the
truth-functional import of a sentence based on pragmatic grounds.  Logical
arguments will of course tend not to have attitudinals in them other than
statements of postulation.

Attitudinals inherently mean that the speaker is emoting when talking,
and the statement is thus to some extent displaced into the world of the
emotional mind (minimal in the case of .oV and .uV attitudinals; the
categorizers may also have significant effect:  "ro'i" almost certainly
throws any attitude into the emotional world, while "ro'a" throws it into
the hypothetical world of social correctness and "re'e" into the realm
of religious faith.)  If you want to ensure that a bridi is unmistakably
taken as a truth claim, you must avoid attitudinals that might throw the
claim out of the real world.  The best way to do this is separation and
repetition, which isn't pragmatically acceptable in highly emotional
situations.  An examples of this, with ".e'a":

do tavla .i .e'anai (go'i)   You're talking (claim) but I forbid this
                             (appropriate to hush someone in a library)
vs.
.e'anai do tavla             (Forbid!) You talk. (ambivalent whether talking
                             actually is happening - perhaps on a sign in the
                             library)
vs.
.e'anai do tavla je'unai     (Forbid!) You talk (non-truth)
                             You aren't talking, but don't start either!
.e'anairo'a.e'aro'i do tavla (Socially forbidden, but emotionally allowed!)
                             You talk.
                             You're talking.  This is impolite and not allowed,
                             but given the current emotional state, we'll
                             tolerate it.

Going back to the original:
>>Until you start sitting up straight and stop playing with your food, young
>>man, there'll be no television for you --- that's for sure!
>
>>He handled this as:
>
>>pu'o le mu'e do co'a xagysirji zutse gi'e na'e kelci le do sanmi doi citno
>>nanmu kei .i'enairo'a do cu .e'anai catlu le se tivni vau ju'o

I believe that the appropriate attitude to translate the English in the
first clause is ".eicai(ro'a)", and the second clause, with negation, is
".aicai".  My version is, therefore:

pu'o le nu do .eicai[ro'a] srasirdra zutse gi'e na'e kelci le do sanmi
kei doi citnau do .aicai na catlu le se tivni

Inchoative of the event of your (You damned well ought to, [it's the
socially correct thing to do]) vertically-straight correctly sitting
and other-than playing with your meal, O young-man, you (Absolutely, as
far as I'm concerned!) are-not:  looker-at the television broadcast.

lojbab