Connectives

[jorge@phyast.pitt.edu]

Here are some comments on the connectives paper.

> If Wishes Were Horses:  The Lojban Connective System
> $Revision: 1.30 $

I like the colourful titles, but this one seems wrong. It seems
to be an example where "if" would not be working as a connective, at least
not as a logical connective.

> 4.  Bridi Logical Connection
...
> This is called the "principle of consistency", and may be paraphrased
> as follows:  "If a false statement is true, any statement follows from it."
> All uses of English "if" must be watched very carefully to see if they
> really fit the Lojban mold.

I really don't like the way this is explained. The sense in which
{ijanai} means "X if Y" is not that X follows from whatever truth
value Y has. 

If Y is true, then it follows that X is true.

If Y is false, then it follows that "X ijanai Y" is true, no matter
whether X is true or not, but X doesn't follow. What does it mean to 
say that "any statement follows from a false statement"? 

  
> 5.  Forethought Bridi Connection
...
> There is a restriction on the bridi used in forethought logical connection:
> the first one cannot have the form of an observative.  That is, there must
> be at least one sumti before the selbri.  See Example 10.9 and following
> text.

(Comment on this later.)

> The syntax of geks is:
> 
> 	[se] GA [nai]

Shouldn't {na ga ... gi ...} be grammatical, to negate the whole sentence? 
 
> 6.  Sumti Connection
...
> Example 6.2 means EXACTLY the same thing as Example 6.1: one may be
> rigorously transformed into the other without any change of logical meaning.
> This rule is true in general for every different kind of logical connection
> in Lojban; all of them, with one exception (see Section 12), can always be
> transformed into a logical connection between sentences expressing the same
> truth function.

Perhaps a comment should be added that the transformation is not always 
a trivial replication of the words. All the examples you give use only 
singular terms, which makes things very simple, but with even a slightly 
more complicated quantifier things get trickier. They can always be 
expanded, but it has to be done with care.

> 8.  Grouping of Afterthought Connectives
...
> But suppose we wished to use afterthought connectives and parentheses.  Two
> pairs of parentheses, analogous to Example 8.6, would seem to be the right
> approach.  But it is a rule of Lojban grammar that a sumti may not begin with
> "ke", so the first set of parentheses must be omitted, producing Example 8.10,
> which is instead parallel to Example 8.7:

If I thought there was any chance for it, I would propose that ke-ke'e be
eliminated for any sumti connection, but I guess it won't happen.
If anybody ever uses them for this purpose, it will be hard to remember
that they can't be used at the start of the sumti. 

> 10.  Multiple Compound Bridi
...
> is the gihek version of Example 8.9.  The same rule about using "ke...ke'e"
> bracketing only just after a connective applies to bridi-tails as to
> sumti, so the first two bridi-tails in Example 10.2 cannot be explicitly
> grouped; implicit left-grouping suffices to associate them.

Again, this rule seems extremely complicated. If ke-ke'e ever come to be
used for this, it will be hard to avoid using them to group the first
terms. Hopefully no one will want to use ke-ke'e for this. Forethought
connectives are much better for grouping, and one has to use forethought
with ke-ke'e anyway, so they seem to be fairly useless for connectors.

> The entire gek-bridi-tail may be negated as a whole by prefixing "na":
> 
> 10.7)	mi na ge klama le zarci gi dzukla le zdani
> 	It is not the case that I both go to the market and walk to the house.

Shouldn't this rule apply also to forethought bridi connection?

	na ge mi klama le zarci gi do dzukla le zdani
 	It is not the case that both I go to the market and you walk to the house.

Is there any parser-reason why this is ungrammatical?

> 10.10)	 ga lo snime cu carvi gi carvi
> 	Either some snow is-raining or [something] is-raining
> 	Either it's snowing or it's raining.
> 
> Unfortunately, the order cannot be reversed:  "*ga carvi gi lo snime cu carvi"
> is ungrammatical, because the parser sees "ga carvi" and assumes bridi-tail
> connection, which the "gi lo snime" thwarts.  

Wouldn't it be possible to treat bridi and bridi-tail connection always 
as the same thing, and thus allow that? That would also allow things like:

	mi ge le zarci cu klama le zdani gi le panka cu dzukla le zdani
	I both to the market go from the house and to the park walk 
	from the house.

But there doesn't seem to be any semantic problem with those. Wouldn't
that simplify things a lot, and at the same time remove the seemingly 
arbitrary restriction on having an observative in first position?

I think the rules would change to something like:

sentence-1<41> = term ... [CU #] bridi-tail | prenex sentence

gek-bridi-tail<54> = gek sentence gik sentence |
	tag KE gek-bridi-tail /KEhE#/ | NA # gek-bridi-tail


This seems like a simplification, making the language more regular and
not having the seemingly arbitrary restriction on "*ga carvi gi lo snime 
cu carvi"

> 14.  Non-Logical Connectives
...
> Example 14.3 covers the case mentioned, where John and Alice divide the
> labor; it also could mean that John did all the hauling and Alice did
> the supervising.  This possibility arises because the properties of a mass
> are the properties of its components, which can lead to apparent
> contradictions: if John is small and Alice is large, then John-and-Alice
> is both small and large.

I'd say John-and-Alice is probably not small in the same context. In any 
case, it will be either small or large as a unit, but not as a direct 
inheritance of John's smallness or Alice's largeness. 

> Single or compound cmavo involving members of selma'o JOI are called joiks,
> by analogy with the names for logical connectives.  It is not grammatical to
> use joiks to connect bridi-tails.

It would really be nice to have gijoiks for that.

> 		A jo'u B	A and B considered jointly

This one is really hard to distinguish from A joi B. In fact, if A jo'u B does
make any sense at all, it is strange that there is no gadri for that:

la djan e la alis    -> le re prenu
la djan joi la alis  -> lei re prenu
la djan ce la alis   -> le'i re prenu

la djan jo'u la alis -> ???


> 14.10)	la djeimyz. jo'u la djordj. remei bruna
> 	James in-common-with George are-a-twosome type-of-brothers.

i la djeimyz jo'u la djordj remei bruna ma?

> The tanru "remei bruna" is not strictly necessary in this sentence,
> but is used to make clear that we are not saying that James and George
> are both brothers of some third person not specified.  

I don't think {remei} avoids anything like that. If George and James are 
Alice's two brothers, then 14.10 would be fine. The way to say that they
are each other's brother would be

	la djeimyz jo'u la djordj bruna simxu

but I don't see the advantage of {jo'u} over {joi} here.


> 16.  Interval Connectives and Forethought Non-Logical Connection
...
> BIhI joiks are grammatical anywhere that other joiks are, including in
> tanru connection and (as ijoiks) between sentences.  No meanings have been
> found for these uses.

How about:

	ta cmalu bi'i barda
	That is somewhere between small and large.

	ta snanu bi'i stici
	That is somewhere between south and west.

	ta verba bi'i makcu
	That is somewhere between a child and an adult.


> 17.  Logical and non-logical connectives within mekso

I suppose that when all the papers are collected in book form, this 
section should be removed from here and sent to the MEX chapter. 
 
> 18.  Tenses, Modals, and Logical Connection
> 
> The tense and modal systems of Lojban interact with the logical connective
> system.  No one paper can explain all of these simultaneously, so each
> paper must present its own view of the area of interaction with emphasis
> on its own concepts and terminology.  For the purposes of this paper,
> the many tenses of various selma'o as well as the modals of selma'o BAI are
> reduced to the simple time cmavo "pu", "ca", and "ba" (of selma'o PU)
> representing the past, the present, and the future respectively.  Preceding
> a selbri, these cmavo state the time when the bridi was, is, or will be true
> (analogous to English verb tenses); preceding a sumti, they state that the
> event of the main bridi is before, simultaneous with, or after the event
> given by the sumti (which is generally a "le nu" abstraction).

This distinction is not really necessary, because when preceding a selbri
these also state that the event of the main bridi is before, simultaneous
with, or after the event of uttering the sentence. {pu broda} is roughly
the same as {broda puku}. 

(Of course, this may not be true for {pu'o} and {ba'o}, but it holds for all 
other tenses and BAIs.)


> 18.6)	mi pu klama le zarci gi'ebabo pu tervecnu lo cidja
> 	I [past] go-to the market, and [later] [past] buy items-of food.

With my proposal (below), we could just use {gibabo} without the logical
connection, which is probably better in general.

Other useful connectors would be the baiboks. Not every BAI gives
a useful connector, but many do. BAIBO for sumti, GIBAIBO for bridi-tail,
IBAIBO for bridi, and BAIGI...GI... for forethought (which already exists,
although this paper doesn't mention them).


> 24.  EBNF for Connectives
> 
> ek<802> = [NA] [SE] A [NAI]
> gihek<818> = [NA] [SE] GIhA [NAI]
> jek<805> = [NA] [SE] JA [NAI]
> joik<806> = [SE] JOI [NAI] | interval | GAhO interval GAhO
> interval<932> = [SE] BIhI [NAI]
> joik-jek<422> = joik # | jek #
> gek<807> = [SE] GA [NAI] # | joik GI # | stag GI #
> guhek<808> = [SE] GUhA [NAI] #
> gik<816> = GI [NAI] #

Proposed additions:

gihek<818> = [NA] [SE] GIhA [NAI] | GI joik-jek

joik-jek<422> = joik # | jek # | stag BO #


Well, that's all for now.

Jorge