Tanru-internal connectives, place structure, non-sense

56 views
Skip to first unread message

iesk

unread,
Oct 31, 2013, 2:25:24 PM10/31/13
to lojban-b...@googlegroups.com
.i je zei .anste

And again I am all of a sudden confused by the simplest Lojbanic expressions. Help!

I understand that, in a tanru of the form {[brodi] brode JA broda}, the place structures of {broda} and {brode} are 'logically connected'. For example, in {ti blanu je zdani} (CLL #12.12), {ti} fills the x1 place of both {blanu} and {zdani}. 

According to CLL, {ti blanu je zdani} means the same as {ti blanu gi'e zdani} (#12.4) and {ti blanu .ije ti zdani} (#12.5). In all three cases, the thing referred to by {ti} both 'blanu's and 'zdani's.

In the above sentences, some sumti places are not filled. The ellipsis of sumti has no effect on the place structure; ellipsised sumti places are understood as {zo'e}. For example, {mi tavla do zo'e zo'e} (CLL #5.5) means the same thing as {mi tavla do} (CLL #5.8)

Likewise, {ti blanu .ije ti zdani} is to be understood as {ti blanu .ije ti zdani zo'e}. And {ti blanu gi'e zdani} is equivalent to {ti blanu gi'e zdani zo'e}.

So far, so good.

However, as for tanru-internal logical connection, {ti blanu je zdani zo'e}, while syntactically correct, is semantically problematic, since nobody knows what it means to {se blanu}: [1(2[zdani1 (home(s)) , blanu1 (blue thing(s)) :] ti this here)2 [is, does] «3(4blanu being blue je and zdani being home(s))4»3 (5[zdani2 (reside-r(s)) , blanu2 (blue??) :] zo'e unspecif it)5]1 (Jbofi'e, for illustration).

So, is {ti blanu je zdani} (CLL #12.12) a meaningful sentence?

To take another example: {sutra je jipci}. Does this express that there is a relation between an obvious x1, which is both quick and a chicken, and an obvious x2, which is both an event quickly brought about by x1 and the chicken-breed of x1—which does not make any sense? Does it mean anything?

I'm afraid that I have probably been using sentences like {sutra je jipci} rather freely, and without bothering to think about any sumti but the x1.

So, what's the solution? (An ellipsised {vau}? If so, then I'll have at least one follow-up question.)

ki'e do mu'o mi'e .iesk.

Michael Turniansky

unread,
Nov 6, 2013, 3:39:20 PM11/6/13
to lojban-b...@googlegroups.com
Yes, if you say "mi sutra je jipci ko'a" (or equivalently, "mi sutra gi'e jipci vau ko'a", then ko'a better darn well refer to some abstraction of something that happens quickly that is also a chicken breed.   This is not likely to be possible in any universe.  However, all is not lost, because, a human will understand that in absence of the the x2(s), the sentence means "mi sutra be zo'e be'o je jipci be zo'e"/"mi sutra be zo'e gi'e jipci be zo'e", so "mi sutra je jipci" is a perfectly fine sentence to say, provided you are a fast chicken, iesk (and a darn intelligent one who can write letters on the lojban list).

      --gejyspa




--
You received this message because you are subscribed to the Google Groups "Lojban Beginners" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lojban-beginne...@googlegroups.com.
To post to this group, send email to lojban-b...@googlegroups.com.
Visit this group at http://groups.google.com/group/lojban-beginners.
For more options, visit https://groups.google.com/groups/opt_out.

iesk

unread,
Nov 8, 2013, 5:36:00 PM11/8/13
to lojban-b...@googlegroups.com

Hm, I don't really {sutra je jipci je mencre} (but thanks for the compliment!), and I'm still not sure what to make of this. But it seems to me that {ti blanu je zdani} is *not* a good example sentence in the reference grammar.


(Since 'the x2 of a one-place brivla' is a separate issue, separately discussed, that need not distract us here, I'll stick to the {sutra je jipci} example.)


la gejyspa ku cusku di'e

> "mi sutra je jipci" is a perfectly fine sentence to say


I agree that 'I am fast, and a chicken' is what that sentence is probably supposed to mean, and what will be understood by a co-operative listener. But then, the speaker should have said {mi sutra gi'e jipci} or so. I don't think {mi sutra je jipci} makes sense. If it did, that would mean that places could be pragmatically cancelled out/filled with {zi'o} … or {no da}, whatever. But I think a legitimate Lojbanic reaction (besides {ki'a}) to {mi sutra je jipci} would always be {sutra je jipci ma}—the second argument is still there and can always be made explicit. Otherwise, we'd end up with place structure ambiguity in ellipsis sentences, wouldn't we?


My tentative conclusion is that CLL #12.12 might better be listed as an erratum. Opinions?

selpa'i

unread,
Nov 8, 2013, 7:04:35 PM11/8/13
to lojban-b...@googlegroups.com
On 08.11.2013 23:36, iesk wrote:
> But it
> seems to me that {ti blanu je zdani} is *not* a good example sentence
> in the reference grammar.

I always treated it like this: When one predicate has more places than
the other, the {je} connection only affects those places that "overlap".
In this example, blanu1 and zdani1 are connected, but zdani2 is not,
because there is no blanu2. Without this, the example would indeed be
questionable, but tanru internal connectives would then be even more
useless, so I don't recommend this interpretation.

> (Since 'the x2 of a one-place brivla' is a separate issue, separately
> discussed, that need not distract us here, I'll stick to the {sutra
> je
> jipci} example.)

Probably a good idea.

> la gejyspa ku cusku di'e
>
>> "mi sutra je jipci" is a perfectly fine sentence to say

I think it's nonsense.

> I agree that 'I am fast, and a chicken' is what that sentence is
> probably supposed to mean, and what will be understood by a
> co-operative listener. But then, the speaker should have said {mi
> sutra gi'e jipci} or so.

Exactly.

> I don't think {mi sutra je jipci} makes
> sense.

Me neither. It would be hard to find anything that both {se sutra} and
{se jipci}.

> If it did, that would mean that places could be pragmatically
> cancelled out/filled with {zi'o} … or {no da}, whatever.

{zi'o} might be debatable, but {no da} is out of the question. There
cannot be a hidden negation in a {zo'e}.

> But I think a
> legitimate Lojbanic reaction (besides {ki'a}) to {mi sutra je jipci}
> would always be {sutra je jipci ma}—the second argument is still
> there
> and can always be made explicit. Otherwise, we'd end up with place
> structure ambiguity in ellipsis sentences, wouldn't we?
>
> My tentative conclusion is that CLL #12.12 might better be listed as
> an erratum. Opinions?

I think it's fine to use tanru internal connectives when all the places
do line up. And I use the rule that "everything lines up with a
non-existent place". My rationale is that the alternative would be
useless.

Also, I think it's important to make a difference between connected
seltau units, and connected tertau units. This thread is only about the
latter. The expansion for connected seltau units is deliberately vague,
and usually gets expanded into two seperate tanru. But this is probably
for another thread.

mu'o mi'e la selpa'i

iesk

unread,
Nov 10, 2013, 11:38:27 AM11/10/13
to lojban-b...@googlegroups.com
la selpa'i ku cusku di'e

>I always treated it like this: When one predicate has more places than
>the other, the {je} connection only affects those places that "overlap".
>In this example, blanu1 and zdani1 are connected, but zdani2 is not,
>because there is no blanu2. Without this, the example would indeed be
>questionable, but tanru internal connectives would then be even more
>useless, so I don't recommend this interpretation.

I understand. You thus save {ti blanu je zdani} from being useless non-
sense. (Even then, {ti blanu je zdani} instead of then-equivalent {ti
blanu gi'e zdani} seems like bad style to me. I think it is a bit of a
pity that it is a CLL example sentence.)

I suppose you also accept {blanu je zdani ti}, by the same reasoning?


>{zi'o} might be debatable, but {no da} is out of the question. There
>cannot be a hidden negation in a {zo'e}.

Indeed.


>Also, I think it's important to make a difference between connected
>seltau units, and connected tertau units. This thread is only about the
>latter. The expansion for connected seltau units is deliberately vague,
>and usually gets expanded into two seperate tanru. But this is probably
>for another thread.

Yes, I decided not to propose the seltau perspective because I find it
easier to explain the issue from the tertau perspective.

Michael Turniansky

unread,
Nov 10, 2013, 5:44:36 PM11/10/13
to lojban-b...@googlegroups.com
  I disagree.  The sentence "mi sutra je jipci" is perfectly sensible, and does  NOT mean the same thing as "mi sutra gi'e jipci" (although ultimately the refferents are the same).   Imagine a Venn diagram.  In one circle are the sutra, fast things (it does not matter what the selsutra are.  They may or may not share them, because they are unspecified.  There are subsets that are sutra be lo nudansu, sutra be lo nu cadzu, etc, but they are ALL included in the "sutra" set.  There is another circle which are jipci (again, some are jipci be la'o gy Rhode Island Red gy, some are jipci be la'o gy Black Shumen  gy, etc.  It doesn't matter.  They are all jipci).  The circles overlap (there are some things that are both sutra je jipci.  I am (well actually, you are, iesk) in that intersection.  That is what the sentence "mi sutra je jipci" represents.  

                   --gejyspa



--

selpa'i

unread,
Nov 10, 2013, 6:02:56 PM11/10/13
to lojban-b...@googlegroups.com
On 10.11.2013 17:38, iesk wrote:
> la selpa'i ku cusku di'e
>>I always treated it like this: When one predicate has more places
>> than
>>the other, the {je} connection only affects those places that
>> "overlap".
>>In this example, blanu1 and zdani1 are connected, but zdani2 is not,
>>because there is no blanu2. Without this, the example would indeed be
>>questionable, but tanru internal connectives would then be even more
>>useless, so I don't recommend this interpretation.
>
> I understand. You thus save {ti blanu je zdani} from being useless
> non-
> sense.

That's right.

> (Even then, {ti blanu je zdani} instead of then-equivalent {ti
> blanu gi'e zdani} seems like bad style to me. I think it is a bit of
> a
> pity that it is a CLL example sentence.)

Could be. It's not easy to find good examples of {broda je brode}.
You'd need something where both predicates have the same valency and
whose places are compatible. The examples tend to be very boring:

(1) ti blanu je xunre
"This is blue and red."

There aren't many unary predicates, and most of them are colors. For
higher arity, the two predicates usually come from the same semantic
family, e.g.:

(2) mi troci je snada lo ka klama
"I tried and succeeded to go."

(even the x3 matches)

Such perfect alignment is rare in Lojban. The place structures just
aren't uniform enough, and often a random "under conditions" or "by
standard" place messes up the symmetry.

> I suppose you also accept {blanu je zdani ti}, by the same reasoning?

Actually, I don't. Here, {ti} fills the x2 of both {blanu} and {zdani},
and since blanu2 doesn't exist, it's a questionable sentence. I see your
point, however.

By the way, I see the same problems with {gi'e} + {vau} connections,
because the places after the "zipper" {vau} also need to match, just
like with a {je} connection. For example:

(3) mi viska gi'e klama vau lo tricu
"I see and-go-to the tree [from some place which is also a
condition for my visual perception]"

So this problem crops up with bridi-tail connection as well.

selpa'i

unread,
Nov 10, 2013, 6:09:28 PM11/10/13
to lojban-b...@googlegroups.com
On 10.11.2013 23:44, Michael Turniansky wrote:
>   I disagree.  The sentence "mi sutra je jipci" is perfectly
> sensible, and does  NOT mean the same thing as "mi sutra gi'e jipci"
> (although ultimately the refferents are the same).   Imagine a Venn
> diagram.  [...] That is what the sentence "mi sutra je jipci"
> represents.  

But if there is nothing that can ever potentially fill the x2 of that
sentence, how can it be sensible?

Michael Turniansky

unread,
Nov 10, 2013, 7:58:13 PM11/10/13
to lojban-b...@googlegroups.com
    Simple... the compound selbri "sutra je jipci" doesn't HAVE an x2 place (but it does have both an s2 and j2 place which have to be internally linked with be), similar to the way that a lujvo may not have all the places that the underlying veljvo seltau jonai tertau have.  

       I don't know why this is a such a problem for you.  This is as specious as saying you can never have a heterosexual woman President of the United States, since then the phrase "first lady" would have no meaning.

                    --gejyspa



--
You received this message because you are subscribed to the Google Groups "Lojban Beginners" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lojban-beginners+unsubscribe@googlegroups.com.
To post to this group, send email to lojban-beginners@googlegroups.com.

selpa'i

unread,
Nov 11, 2013, 9:21:30 AM11/11/13
to lojban-b...@googlegroups.com
On 11.11.2013 01:58, Michael Turniansky wrote:
>     Simple... the compound selbri "sutra je jipci" doesn't HAVE an x2
> place (but it does have both an s2 and j2 place which have to be
> internally linked with be), similar to the way that a lujvo may not
> have all the places that the underlying veljvo seltau jonai tertau
> have.  

Okay, now I at least understand what you are saying. I find this
problematic, as tanru are by their very nature ad-hoc, and figuring out
such eliminated places on the fly is likely impossible (assuming for a
second that it's sensible).

If you say it has no x2, then I assume you would say that in:

mi klama je sutra lo tricu

{lo tricu} actually fills klama3 and it fills neither klama2 nor
sutra2. How can this be practically used? In which case it would mean:

mi klama je sutra lo tricu
"I go [and-am-fast] from the tree."

Is this not weird?

>        I don't know why this is a such a problem for you.  

Apart from the aforementioned practical problems your idea presents, I
think it's also strange for places to disappear from a *tanru*. In my
opinion, when you join two things with {je} (in this case the sumti
places), then you claim that they both apply, not that the places
disappear when nothing can satisfy the claim. It seems somewhat
backwards.

Michael Turniansky

unread,
Nov 11, 2013, 12:04:31 PM11/11/13
to lojban-b...@googlegroups.com
On Mon, Nov 11, 2013 at 9:21 AM, selpa'i <m...@plasmatix.com> wrote:
On 11.11.2013 01:58, Michael Turniansky wrote:
    Simple... the compound selbri "sutra je jipci" doesn't HAVE an x2
place (but it does have both an s2 and j2 place which have to be
internally linked with be), similar to the way that a lujvo may not
have all the places that the underlying veljvo seltau jonai tertau
have.  

Okay, now I at least understand what you are saying. I find this problematic, as tanru are by their very nature ad-hoc, and figuring out such eliminated places on the fly is likely impossible (assuming for a second that it's sensible).

If you say it has no x2, then I assume you would say that in:

   mi klama je sutra lo tricu

{lo tricu} actually fills klama3 and it fills neither klama2 nor sutra2. How can this be practically used? In which case it would mean:

 
  No, I would not say that.  When I said " 'sutra je jipci' doesn't have an x2," I wasn't generalizing to _all_ broda JA brode.  If you SPECIFY something in that spot, as in your sentence, you ARE saying the same as "mi klama gi'e sutra vau lo tricu" and you darn well better be using something that can fit both of the underlying x2 places (Which obviously "lo tricu" cannot) .  What I was asserting was that since the set of all things that are both se sutra and se jipci is an empty set (as far as I can reasonably ascertain), it /de facto/ has no x2 place, not that it has no x2 place by nature of the construction.

  Therefore as long as your sentence doesn't try to put something into that place, there is no problem with asserting the sentence "mi sutra je jipci" is meaningful (I guess, to put it another way, I am asserting that an implicit "zo'e" (but probably not, pe'i, an explicit "zo'e") in fact doesn't have to be something that can really exist.  (kind of like the way that mathematicians 400 years ago would all agree that the square root of a negative number is complete and utter rubbish, and any student who suggested otherwise should be beaten with a switch until he learned the proper ways of mathematics, but goshdarnit,. don't all these equations and proofs  work out nicely if we PRETEND there could be such a thing?)
 
   mi klama je sutra lo tricu
   "I go [and-am-fast] from the tree."

Is this not weird?


       I don't know why this is a such a problem for you.  

Apart from the aforementioned practical problems your idea presents, I think it's also strange for places to disappear from a *tanru*. In my opinion, when you join two things with {je} (in this case the sumti places), then you claim that they both apply, not that the places disappear when nothing can satisfy the claim. It seems somewhat backwards.


mu'o mi'e la selpa'i


    --gejyspa
 

selpa'i

unread,
Nov 11, 2013, 4:22:30 PM11/11/13
to lojban-b...@googlegroups.com
On 11.11.2013 18:04, Michael Turniansky wrote:
> On Mon, Nov 11, 2013 at 9:21 AM, selpa'i <m...@plasmatix.com> wrote:
>> If you say it has no x2, then I assume you would say that in:
>>
>>    mi klama je sutra lo tricu
>>
>> {lo tricu} actually fills klama3 and it fills neither klama2 nor
>> sutra2. How can this be practically used? In which case it would mean:
>
>   No, I would not say that.  When I said " 'sutra je jipci' doesn't
> have an x2," I wasn't generalizing to _all_ broda JA brode.  If you
> SPECIFY something in that spot, as in your sentence, you ARE saying
> the same as "mi klama gi'e sutra vau lo tricu" and you darn well
> better be using something that can fit both of the underlying x2
> places (Which obviously "lo tricu" cannot) .  

But there is something in that place either way. It's either something
explicit, or it's {zo'e}.

> What I was asserting was
> that since the set of all things that are both se sutra and se jipci
> is an empty set (as far as I can reasonably ascertain), it /de facto/
> has no x2 place, not that it has no x2 place by nature of the
> construction.

Okay. In other words, there is an x2, but nothing can sensibly fill it.
The moment you use it, the sentence becomes nonsense.

{.i sutra je jipci} is the same as {.i zo'e sutra je jipci zo'e}. It
claims that something fills that x2.

>   Therefore as long as your sentence doesn't try to put something
> into that place, there is no problem with asserting the sentence "mi
> sutra je jipci" is meaningful (I guess, to put it another way, I am
> asserting that an implicit "zo'e" (but probably not, pe'i, an
> explicit
> "zo'e") in fact doesn't have to be something that can really exist.

A {zo'e} always has some value that makes the bridi true. But whatever
the {zo'e} in {sutra je jipci}'s x2 is, it cannot make the bridi true
(because no such value exists), so the whole predicate is faulty.

Michael Turniansky

unread,
Nov 11, 2013, 5:12:50 PM11/11/13
to lojban-b...@googlegroups.com
(arrghh.. this is getting far beyond beginner level)
  
  But I claim that an implicit zo'e is NOT the same as an explicit one (I will leave open the question of whether you can you can use the value of "zi'o" for a "zo'e"  I don't think you can, because it creates an entirely new selbri, but I will  be willing to accept the opposite argument)... and I think I can get you to agree.  Consider the following...

  1)   Some selbri have INFINITELY many numbered places (du and jutsi, notably) So are you asserting that if something doesn't have a higher order in a taxonomy, for example, that I have to add an infinite number of "zi'o" to any sentence using "jutsi" (since "zo'e" won't do?).

  2) (weaker argument)  Besides numbered places, every selbri has a infinite number of unnumbered places, from BAI and all FIhO SELBRI FEhU. (and reasonably, other sumti tcita such as tenses might be considered to be subsumed in here as well).  Do all these have implicit zo'e?  How about those that cannot make any sense?  (You can easily undercut this argument by saying "implicit zo'e-dom only applies to numbered places.  But then again, remember that you CAN say "mi blanu lo mlatu" and there "lo mlatu" is basically equivalent to "do'e lo mlatu", but it is explicitly in a numbered (but unknown relationship) sumti place.  You can have an infinite number of these, so this becomes argument #1).

  If you accept either argument 1 or 2, you must concede that either every sentence is "faulty" (your word), or that you don't have to specify every place, and that doesn't invalidate bridi (so implicit zo'e must either be different than explicit zo'e, or can also include values like "zi'o" or "no da")

                     --gejyspa






Jacob Errington

unread,
Nov 11, 2013, 5:52:45 PM11/11/13
to lojban-b...@googlegroups.com
Not really sure who to quote at this point, so I'll just talk in general.

My opinion on the business of empty places and zo'e is that an empty place is not strictly equivalent to an overt {zo'e}. First of all, the idea that an empty place always equates to an overt {zo'e} (minus place structure advancement) falls apart in relative clauses and {ka}-abstractions, where implicit {ke'a} and {ce'u} are crucial to brevity. I much prefer the idea that if a place is left empty, then it could very well be {zi'o} due to the underlying idea that I subscribe to such that {ko'a broda} implies {zi'o broda}.

So an empty place can give rise to {zo'e} and {zi'o} ({zu'i} is just a special kind of {zo'e} IMO) and in certain clauses, the empty place can also give rise to {ce'u} and {ke'a}. Furthermore, I believe that it is in some sense an axiom of Lojban that implicit information cannot produce a contradiction, va'i that an empty place can give rise to {no da}, since that would result in a contradictory negation.

As for the problem of the place structure of a tanru whose tertau is a JA-connected selbri, I think that the most sensible solution is the toss our hands into the air and let pragmatics decide. If one says {mi jipci je sutra lo ka tavla}, then it would be silly for one to assume that {lo ka tavla} is equally filling jipci2. That's just nonsense, and I think that we should try to limit nonsensical interpretations as much as possible. If you want to explicitly make all the places line up, then do so using giheks; they can do that quite well, and no one really argues about how to interpret {broda gi'e brode vau ko'a}. Yes, it's longer, but that's the price you pay for concision. (I think of that as the length-complexity axiom: precision and length should be proportional as best as possible.) So I think that in the general case of {broda je brode ko'a}, the most sensible thing to say is that we don't know the structure of the trailing sumti in such a case, in much the same way that we don't know the distributivity of a bare lo-sumti. We are thus complying with the length-complexity axiom by making a very short construct such as a je-connective in a tertau have a very vague meaning. The listener is expected to infer the most sensible structure based on context. On the other hand, it is pretty useful to require the x1 of the left and right operands of the connective to line up, and is a very weak requirement, much less weak than that all the places line up, which, I agree, is just plain silly, bordering on totally useless in practice.

.i mi'e la tsani mu'o

iesk

unread,
Dec 23, 2015, 2:04:28 PM12/23/15
to Lojban Beginners
Browsing through the example sentences of https://mw.lojban.org/papri/The_Crash_Course_(a_draft) reminds me of this issue. I really hate this. JA is so useless.

Sorry, just letting off steam. ju'i nai
Reply all
Reply to author
Forward
0 new messages