searching

[Luke Bergen]

Going through the gismu on memrise (awesome service by the way), and I got to "sisku".

Apparently I've been using this really core word completely wrong.  I've been using it like "mi sisku lo mi ckiku", when evidently x2 is a property and x3 is a set.... of somethings?

I'm kind of confused by this definition.  So, if there is a set of properties, and there is one property in that set which x1 seeks among that set x3.

I'm just a little confused about how this is meant to be used.

Meanwhile, what do I say when I'm looking for my car keys?  "mi troci co facki lo ckiku"?

Except, double checking to make sure that I've got that right, even that seems wrong since to "facki" is to discover a fact about a subject/object.  I guess an object's location is a fact about it.... but still.... I just want to find my damn car keys, how do I say that?!

[Ian Johnson]

As defined, sisku means "x1 looks among the members of set x3 for something which has property x2". To clarify, it does not mean that you're looking for a particular object, which is why a property is used; I can be looking for "something green" without caring which green thing I ultimately find.

This definition is rather cumbersome, and especially rather frustrating in situations like the one you describe where you're looking for a concrete object, which is of course the most common situation.

That said, {facki} and {troci co facki} can work using {fi}, because when you find your keys you {facki lo du'u lo ckiku cu zvati ma kau kei lo ckiku}. Of course, {troci co facki fi lo ckiku} is rather frustratingly vague, while {troci co facki lo ckiku cu zvati ma kau} is frustratingly verbose.

mi'e la latro'a mu'o

--
You received this message because you are subscribed to the Google Groups "lojban" group.
To post to this group, send email to loj...@googlegroups.com.
To unsubscribe from this group, send email to lojban+un...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/lojban?hl=en.

[Stela Selckiku]

On Tue, Jan 8, 2013 at 11:50 AM, Luke Bergen <lukea...@gmail.com> wrote:

> I guess an object's location is a fact about it.... but still.... I just
> want to find my damn car keys, how do I say that?!

coi la pafcribe coi pendo .i mi ta'e pilno zo zvafa'i .i se'o lo lujvo
cu cafne za'e badyxau

mi'e la stela selckiku

mu'o

[Ian Johnson]

The second one was supposed to be {troci co facki lo du'u lo ckiku cu zvati ma kau}, .u'u

mi'e la latro'a mu'o

[Jacob Errington]

Using Lojban as it is without changing anything, there is a convoluted
way to use sisku as-is, like this:
.i mi sisku lo ka lo mi ckiku cu zvati ce'u kei ro zo'e ne lo zdani
This happens to make sisku actually means something more along the
lines of "x1 checks if the function x2 applies to members of set x3".
However, this means you can say {.i mi sisku tu'a lo mi ckiku} and
leave out the x3, which is still a bit weird.
There has been talk about chaning sisku to mean something like "x1
looks for object x2 in/at location(s) x3", but it's doubtful that a
gismu will get changed.
Otherwise, you were on the right track with troci+facki, but facki2 is
indeed a proposition, but we can use an indirect question to help us
out:
.i mi troci lo ka facki lo du'u makau se zvati lo mi ckiku
This form is fully compositional and can be made into a 100%* jvajvo:
{selzvafaktoi}

So in sum, if you want to keep sisku as it is (which isn't really so
useful) you can say {.i mi sisku tu'a lo mi ckiku}.
If you want sisku to change, use it with its alternate definition: {.i
mi sisku lo mi ckiku}.
If you want to use a lujvo to avoid sisku altogether, use {.i mi
selzvafaktoi lo mi ckiku}.

.i mi'e la tsani mu'o

[Jacob Errington]

Little correction:

On 8 January 2013 12:08, Jacob Errington <nict...@gmail.com> wrote:
> Using Lojban as it is without changing anything, there is a convoluted
> way to use sisku as-is, like this:
> .i mi sisku lo ka lo mi ckiku cu zvati ce'u kei ro zo'e ne lo zdani

{ro zo'e ne lo mi zdani} distributes, which is bad, so we need to make
it into a mass/set:
{loi ro zo'e ne lo mi zdani} is better, but you could probably use {lo
zdani} outright as a sisku3, in which case you're considering the
house to be a mass/set of all the places within (because xorlo).

[Luke Bergen]

Yeah, I wasn't looking for a "we should change lojban suck that..." sort of thing.  I assumed there was some solution.

Just kind of sucks that it's very easy for me to say "I am looking for a green thing in my dresser" (mi sisku lo (ka?) crino lo seldacru).  But for "I'm looking for my keys", I've either got to spend half the day spelling it all out, or I guess selkik's tanru works  tolerably well "mi troci co zvafa'i lo mi ckiku"

I guess it just seems like an unfortunate definition for sisku given how easy it is to add an object-by-property aspect to the searching.  i.e.  "mi sisku lo creka poi crino" or even "mi sisku lo crino creka".

meh, it is what it is.

[Luke Bergen]

lo'ai suck sa'ai such le'ai

[Ian Johnson]

If you defined it the other way, you wouldn't get quite the same thing by just attaching {poi}. Instead you could wind up looking for a definite shirt, rather than just "some green shirt". {da} wouldn't fix this either, because the quantifier would be in the wrong place. (That is, "I'm looking for {da poi crino}" is "There exists a green thing that I'm looking for"; the existential is outside the looking. To move it inside you would need an abstractor, which is exactly what gimste-sisku does.) 

Ultimately I agree, {sisku} as currently defined is rather strange to consider as the primitive. It feels excessively abstracted from everyday life, even if it is occasionally rather compact/precise.

mi'e la latro'a mu'o

On Tue, Jan 8, 2013 at 12:25 PM, Luke Bergen <lukea...@gmail.com> wrote:

[John E Clifford]

This points to the reason for the "odd" place structure.  If I say {mi sisku lo pavelseljirna}, it follows that I am claiming that there is a unicorn I am looking for, and hence that there is a unicorn.  Which is not what is intended at all.  So, in this case and a few others (not nearly enough to prevent problems, but enough to prevent looking for a good general solution), places were altered to give non-problematic places, at the cost of practical usage.  The solution to allow just different kinds of sumti in places was never developed as much as it needed to be, although a lot of work was done with {djica} a few other problem cases.

---

From: Ian Johnson <blindb...@gmail.com>
To: loj...@googlegroups.com
Sent: Tuesday, January 8, 2013 11:38 AM
Subject: Re: [lojban] searching

[selpa'i]

la'o me. Luke Bergen .me cusku di'e

> Yeah, I wasn't looking for a "we should change lojban suck that..." sort
> of thing. I assumed there was some solution.
>
> Just kind of sucks that it's very easy for me to say "I am looking for a
> green thing in my dresser" (mi sisku lo (ka?) crino lo seldacru). But
> for "I'm looking for my keys", I've either got to spend half the day
> spelling it all out, or I guess selkik's tanru works tolerably well "mi
> troci co zvafa'i lo mi ckiku"

With sisku = "x1 looks for object x2 at location x3", you can still say
you're looking for a green thing:

mi sisku lo crino

The property is not very convenient, so I always use sisku with the
above definition.

> I guess it just seems like an unfortunate definition for sisku given how

> easy it is to /add/ an object-by-property aspect to the searching. i.e.

> "mi sisku lo creka poi crino" or even "mi sisku lo crino creka".

I would say the above sisku is the most basic predicate, and the old
sisku can be retrieved with something like "kairsisku".

mu'o mi'e la selpa'i

[selpa'i]

la. pycyn. cu cusku di'e

> This points to the reason for the "odd" place structure. If I say {mi
> sisku lo pavelseljirna}, it follows that I am claiming that there is a
> unicorn I am looking for, and hence that there is a unicorn. Which is
> not what is intended at all.

With old lo this was the case, but isn't anymore with xorlo. "lo
pavyseljirna" is Unicorn, but there is no existential import. You can
search for "lo pavyseljirna" even if none exist.

[selpa'i]

la latro'a cu cusku di'e

> If you defined it the other way, you wouldn't get quite the same thing
> by just attaching {poi}. Instead you could wind up looking for a
> definite shirt, rather than just "some green shirt".

Xorlo "lo creka" does not refer to a specific shirt, that would be "le
creka", so there is no problem with using lo here, but:

> {da} wouldn't fix
> this either, because the quantifier would be in the wrong place. (That
> is, "I'm looking for {da poi crino}" is "There exists a green thing that
> I'm looking for"; the existential is outside the looking.

Right, "da" doesn't work for these situations, that's one of the
problems xorlo fixed.

[John E Clifford]

Always could; existence is a predicate, being there is the UD.  So, {lo pavyseljirna cu broda} always implies {da pavyseljirna } but never (directly) {da poi pavyseljirna cu zasti}.

---

From: selpa'i <sel...@gmx.de>
To: loj...@googlegroups.com
Sent: Tuesday, January 8, 2013 1:24 PM
Subject: Re: [lojban] searching

-- You received this message because you are subscribed to the Google Groups "lojban" group.
To post to this group, send email to loj...@googlegroups.com.

To unsubscribe from this group, send email to lojban+unsub...@googlegroups.com.

[v4hn]

On Tue, Jan 08, 2013 at 11:31:21AM -0800, John E Clifford wrote:
> Always could; existence is a predicate, being there is the UD.
> So, {lo pavyseljirna cu broda} always implies {da pavyseljirna}
> but never (directly) {da poi pavyseljirna cu zasti}.

Could you please elaborate on that a bit?
Neither do I agree that existence is a predicate(not on the level
of semantical analysis), nor do I agree with the implications
you derive from {lo brode cu broda}.
But maybe I just didn't get you...


v4hn

[selpa'i]

la .pycyn. cu cusku di'e

> Always could; existence is a predicate, being there is the UD. So, {lo
> pavyseljirna cu broda} always implies {da pavyseljirna } but never
> (directly) {da poi pavyseljirna cu zasti}.

{lo pavyseljirna cu broda} does not imply {da pavyseljirna}, but {da
pavyseljirna} implies {lo pavyseljirna cu broda}. The universe of
discourse is not the only thing of importance here. Let's say I'm a
hunter, and it so happens that the species I'm hunting for has gone
extinct earlier that day (but I don't know that). For simplicity sake,
let us say that it was deer. There are no deer left.

In this scenario, we can make certain claims about lo mirli, without
implying anything about its existence:

mi sisku lo mirli
I'm looking for deer.

Here, {lo mirli cu broda} but {na ku da mirli}.

[John E Clifford]

Basic logic: KFaGa => SxFx
Existence is a predicate in Lojban (and arguably in English, etc.), this is why S or whatever you use is properly called an particular quantifier, not an existential one. 
There *may* be some conflict between log and lang here, as there often is, but the log side is pretty clear.

---

From: v4hn <m...@v4hn.de>
To: loj...@googlegroups.com
Sent: Tuesday, January 8, 2013 4:54 PM
Subject: Re: [lojban] searching

[Ian Johnson]

It's not that it doesn't, only that it needn't. Specifying explicitly that it doesn't isn't possible. va'i xorlo doesn't really "fix" this per se, in that under the modified "x1 looks for x2 at x3" definition, you can't explicitly say "I'm looking for a green thing, I don't care which one", at least not without introducing another bridi.

mi'e la latro'a mu'o

--
You received this message because you are subscribed to the Google Groups "lojban" group.
To post to this group, send email to loj...@googlegroups.com.

To unsubscribe from this group, send email to lojban+unsubscribe@googlegroups.com.

[Ian Johnson]

His point involves a definition of {da broda} as "there exists x in the universe of discourse such that x broda". This does not require "There exists x in the universe of discourse such that x broda gi'e zasti [fi zu'i]", since we talk about things that don't exist all the time. Under this definition, if you say {lo pavyseljirna}, there's a unicorn in the universe of discourse, irrespective of whether it exists in reality.

mi'e la latro'a mu'o

--
You received this message because you are subscribed to the Google Groups "lojban" group.
To post to this group, send email to loj...@googlegroups.com.

To unsubscribe from this group, send email to lojban+unsubscribe@googlegroups.com.

[v4hn]

On Tue, Jan 08, 2013 at 06:56:09PM -0500, Ian Johnson wrote:
> It's not that it doesn't, only that it needn't. Specifying explicitly that
> it doesn't isn't possible. va'i xorlo doesn't really "fix" this per se, in
> that under the modified "x1 looks for x2 at x3" definition, you can't
> explicitly say "I'm looking for a green thing, I don't care which one", at
> least not without introducing another bridi.

... And we are back to the "any" discussion.
This, however, has not really anything to do with {sisku} /per se/.

I agree with selpa'i in that it should work out quite well to use lo(also le
for e.g. "my keys") here (using the modified definition) and that this change
makes a lot of sense due to xorlo.


v4hn

[selpa'i]

la latro'a cu cusku di'e

> His point involves a definition of {da broda} as "there exists x in the
> universe of discourse such that x broda". This does not require "There
> exists x in the universe of discourse such that x broda gi'e zasti [fi
> zu'i]", since we talk about things that don't exist all the time. Under
> this definition, if you say {lo pavyseljirna}, there's a unicorn in the
> universe of discourse, irrespective of whether it exists in reality.

"Reality" (which you described with zasti fi zu'i) doesn't matter, since
we're talking in terms of UDs. lo pavyseljirna introduces unicorns into
our UD, but it still doesn't claim that one *exists* in *this UD* (other
UDs don't matter, nor does "reality"). This distinction seems important,
and it matters a lot when we talk about xorlo too. So, I'm not sure this
is exactly the point he makes. Some clarification would be useful.

[selpa'i]

la latro'a cu cusku di'e

> It's not that it doesn't, only that it needn't. Specifying explicitly
> that it doesn't isn't possible.

Why not?

> va'i xorlo doesn't really "fix" this per
> se, in that under the modified "x1 looks for x2 at x3" definition, you
> can't explicitly say "I'm looking for a green thing, I don't care which
> one", at least not without introducing another bridi.

Why? What is the problem with "mi sisku lo crino"? This is exactly the
kind of thing xorlo fixes.

[Felipe Gonçalves Assis]

I am pretty sure that {lo broda cu brode} implies {da poi broda cu brode}
(except that {da} should be a plural variable). In particular, {lo broda}
presupposes {da broda}.

When I say {mi sisku lo crino}, there is no green thing, and no green bunch
of things out of which I can create meaning for the "looking for" predicate.
You could argue that I am looking for something that is among all the green
things, {mi sisku lo ka me lo [ro] crino}, which is ok, except for the problem
with non-existent objects. If I say {mi sisku [lo ka me] lo pavyseljirna}, I am
making a commitment with the belief that da pavyseljirna. It can't be just
an imaginary unicorn, because I am not looking for a unicorn on my mind.
You can also consider contradictory properties. If you said {mi sisku lo blanu
poi na blanu}, I would reply {na'i go'i}! ({mi sisku lo ka blanu gi'e na blanu}
is fine, though).

The non-existence thing could be solved by working with sets, but it is
logically cleaner to work directly with their defining properties.

mu'o
mi'e .asiz.

> --
> You received this message because you are subscribed to the Google Groups
> "lojban" group.
> To post to this group, send email to loj...@googlegroups.com.
> To unsubscribe from this group, send email to

> lojban+un...@googlegroups.com.

[John E Clifford]

Of course it does, KFaFa => SxFx  and any variation of that you want to use to cover the Lojban will work essentially the same.  And, clearly, Fa=/=> KFaGa, and particularly not for G = exists.  And, of course, again, even if deer no longer exist they may be in the UD (and surely are if I assert {mi sisku lo mirli} in the thing sense of {sisku}) and so there are some, just not any longer in the realm of existents.  We talk about (have in our UD) non-existent things all the time (Sherlock Holmes, unicorns, caring Republicans, fiscally sound Democrats, etc.) and have no problems with them.

---

From: selpa'i <sel...@gmx.de>
To: loj...@googlegroups.com

Sent: Tuesday, January 8, 2013 5:39 PM
Subject: Re: [lojban] searching

[John E Clifford]

Careful:  "imaginary" is often contrasted with "existent", so that is not the relevant distinction.  Unicorns, when in the UD, are also in the extension of "imaginary".  The real problem is that I can search for things that are not even in the UD (impossible things may be a case, but I don't want to exclude them prematurely).  Also, imaginary things are not in your mind, at least as aplace to look fro them (but this gets into the messiness of (real) epistemology and representation theory, which seems wise to avoid now). 
I'm not clear howsets solves working with non-existent things -- even C-sets don't seem to help.

---

From: Felipe Gonçalves Assis <felipe...@gmail.com>
To: loj...@googlegroups.com
Sent: Tuesday, January 8, 2013 6:58 PM
Subject: Re: [lojban] searching

> lojban+unsub...@googlegroups.com.

> For more options, visit this group at
> http://groups.google.com/group/lojban?hl=en.
>

--
You received this message because you are subscribed to the Google Groups "lojban" group.
To post to this group, send email to loj...@googlegroups.com.

To unsubscribe from this group, send email to lojban+unsub...@googlegroups.com.

[John E Clifford]

What if the extension of {crino} is null?  Of course it can't be if {mi sisku lo crino} is accepted in the proposed sense, but that is questionbegging.

---

From: selpa'i <sel...@gmx.de>

To: loj...@googlegroups.com
Sent: Tuesday, January 8, 2013 6:20 PM
Subject: Re: [lojban] searching

-- You received this message because you are subscribed to the Google Groups "lojban" group.
To post to this group, send email to loj...@googlegroups.com.

To unsubscribe from this group, send email to lojban+unsub...@googlegroups.com.

[Ian Johnson]

Because {lo crino} can very well mean "that blade of grass over there". It's not specific either way, and unlike with definiteness there isn't a way to add quantifiers or anything to make it so.

mi'e la latro'a mu'o

[Ian Johnson]

Huh? If it's in the UD, it exists in the sense of the UD in the UD. That is, a UD induces a zasti3 as "x exists iff x is in the UD". That's definitions, not even logic. I think what you're saying is that under xorlo introducing {lo pavyseljirna} into the UD doesn't imply that {da pavyseljirna} in the sense of the UD. But {lo pavyseljirna} still exists in the sense of the UD, whatever that means, and more likely than not {da pavyseljrina} in the sense of the UD (if it doesn't, then you're being a jerk, in that what you're saying is basically a bear goo example). 

We really need to have a chat about how to better handle universes of discourse, I think...

mi'e la latro'a mu'o

[v4hn]

On Wed, Jan 09, 2013 at 02:10:28AM -0500, Ian Johnson wrote:
> Because {lo crino} can very well mean "that blade of grass over there".

What's the problem here? If you're just looking for something green
and your universe of discourse includes blades of grass, then this is
perfectly valid.

> It's not specific either way, and unlike with definiteness there isn't a
> way to add quantifiers or anything to make it so.

I don't get the rest.. If you are looking for a green shirt
and your UD includes more things than shirts, you can always say
{mi sisku lo crino creka}


v4hn

[v4hn]

On Tue, Jan 08, 2013 at 06:48:04PM -0800, John E Clifford wrote:
> Of course it does, KFaFa => SxFx [...]

Please explain your notation.


v4hn

[selpa'i]

la .pycyn. cu cusku di'e

> What if the extension of {crino} is null? Of course it can't be if {mi
> sisku lo crino} is accepted in the proposed sense, but that is
> questionbegging.

I don't think we're making claims about the extension of {crino} when
saying {lo crino}. lo is an intensional article. The deer example makes
this quite clear, we can be looking for deer even if none exist. So
there is no problem if the extension is null.

[John E Clifford]

Oh dear, is this one going around again?  {lo} is not intensional (whatever that means exactly), even in xorlo (except when xorxes go off on a tangent). {lo broda} refers to brodas, not brodaness or Mr. Broda or whatever, and brodas in the present UD, not in cloud-cuckooland.  Existence, unmodified, refers to existence in the putative "real world", a subset of the UD for logical purposes.  If the extension of {broda} is null in the UD, then {lo broda} is improper, although, pragmatically, using it enlarges the UD to include at least one broda (assuming that the expression is allowed, UDs are collaborative, after all).

---

From: selpa'i <sel...@gmx.de>
To: loj...@googlegroups.com

Sent: Wednesday, January 9, 2013 8:57 AM
Subject: Re: [lojban] searching

[Ian Johnson]

Look at the "any" discussion, as it's the same idea. {lo crino} can very well have only one referent, even if there are many green things in the UD. It shouldn't have only one referent, if the distinction matters to you as the speaker; that's (part of?) what {le} is for. But it can anyway. This means you can't say "I'm looking for a green thing" and be absolutely certain that your listener knows you don't care which one (without a separate bridi).

mi'e la latro'a mu'o

[Ian Johnson]

Also, note that the solution to some of the problems from the "any" discussion, namely an outer quantifier of {pa}, doesn't work here. For example, in {ko dunda pa lo plise mi}, the command is satisfied no matter which apple among {lo plise} the listener gives, so it is a satisfactory translation of "Give me any one of the apples". This works in the {ko} case because the {ko} creates a scope that the {pa} quantifier is under, which is the origin of the awkward "make it true that" translations of Lojban imperatives to English. (This is not a bad thing, and is arguably a good thing.) Without a {ko} this requires some kind of abstraction or similar. 

mi'e la latro'a mu'o

[John E Clifford]

Well, for indefiniteness you can (and should) use variable rather than {lo, le} both of which involve saliency (external and internal), which is precisely something definite, relatively speaking.

---

From: Ian Johnson <blindb...@gmail.com>
To: loj...@googlegroups.com

Sent: Thursday, January 10, 2013 9:09 AM
Subject: Re: [lojban] searching

--
You received this message because you are subscribed to the Google Groups "lojban" group.
To post to this group, send email to loj...@googlegroups.com.

To unsubscribe from this group, send email to lojban+un...@googlegroups.com.

[Ian Johnson]

As I said in response to the remark you quoted, indefiniteness doesn't fix searching, because you need a scope around the quantifier that you don't have without an abstractor. {mi sisku da poi crino} is "There exists a green thing for which I am searching", which means that "within the searching" if you will there is a definite green thing.

mi'e la latro'a mu'o

[John E. Clifford]

Right.  But bare {lo/le crino} doesn't work either, because of salience.  Hence the odd place structure of {sisku}.

Sent from my iPad

[v4hn]

Yes, we still don't have an "any" in lojban.
We discussed that for the last 3 weeks.

Still, I don't see that {mi sisku lo crino} fails to mean the right thing
- given that we don't have an "any". If you think that is the problem, go
back to the last discussion and continue it -
It means something like "I'm looking for something that is green".

Talking about existence again. The {lo crino} introduces {da poi crino}
into the context. This, however, does in my opinion not need to be a
specific _physical_ item at this moment. The {da} is one object in the
universe of discourse, which _only_ has the primitive property {da crino} for now.
This object in the UD does not have to map to a physical object(talking about unicorns)
and might also have multiple physical objects it _can_ map to.
Looking for (any)something green, you _do_ have an object, in the sense
of an element in the universe of discourse, in mind. Namely,
an item, whos only relevant property is being green.
If someone points you to a blade of grass and you are not satified with it,
then you have additional constraints which you didn't mention up to then.
Still, you can go on and specify {go'i fe lo crino creka} or something like that,
adding more specific properties to the object in the UD and therefore
restricting the possible physical objects it can refer to.

If you have a specific item in mind, on the other hand, surly you can
just say {mi sisku le crino} and there are no problems with that.

On the basis of this interpretation, it seems to work out well,
to say {mi sisku lo crino} if you mean to say that you are looking
for something green.

{ma pinka}


v4hn

On Thu, Jan 10, 2013 at 10:09:04AM -0500, Ian Johnson wrote:
> Look at the "any" discussion, as it's the same idea. {lo crino} can very
> well have only one referent, even if there are many green things in the UD.

> It *shouldn't *have only one referent, if the distinction matters to you as

[John E Clifford]

Howsabout going back to the basics of "any" in English?  It is a context leaper, a universal embedded in a verso context with scope over the whole in which the context is subordinate.  So, what we want is Ax Greenx I seek x. Not, notice, {mi sisku ro crino}, because {sisku}(in the thing sense, not the property sense) is short for "has a goal which would be fulfilled if I were to have (in whatever the appropriate sense is) x" and so every green thing fits and none is special ("if my goal were fulfilled, I would have").  This still supposes, of course, that there is something green in the UD, so the property sense is still better.  Of course, spelling out the counterfactual stuff in such a way as to make the quantifier scope points clearer would be nice, too, but no one seems to like {tu'a} and it is a little iffy around the edges anyhow.

---

From: v4hn <m...@v4hn.de>
To: loj...@googlegroups.com
Sent: Thursday, January 10, 2013 6:19 PM
Subject: Re: [lojban] searching

[v4hn]

On Thu, Jan 10, 2013 at 06:42:53PM -0800, John E Clifford wrote:
> Howsabout going back to the basics of "any" in English?
>
> It is a context leaper, a universal embedded in a verso context
> with scope over the whole in which the context is subordinate.

> So, what we want is Ax Greenx I seek x.

I already asked you two days ago to explain your notation, please.
Does this mean something like "forall x : Green(x).(seek(I, x))"
or "A(x) => Green(x) => seek(I, x)" or "forall x. Green(x) => seek(I, x)"
...

> Not, notice, {mi sisku ro crino}, because {sisku}(in the thing sense,
> not the property sense) is short for "has a goal which would be fulfilled
> if I were to have (in whatever the appropriate sense is) x" and so every
> green thing fits and none is special ("if my goal were fulfilled, I would have").

I'm not sure I like that "goal driven" analysis of seek.
Especially, mixing up quantifiers and goal constraints is rather confusing.

What your "Ax Greenx I seek x" is _supposed_ to mean, I think, is the following.

There exists a goal G1 which I have in mind, such that for all green things
it is true that if I have such a green object, the goal G1 is fulfilled.

I very much prefer the analysis I described in my last mail,
because if you try to apply quantifiers here, you have to be explicit about
the existential goal quantification.
Else you could end up searching for _all_ green things:

forall x. there exists g. have(I, x) => satisfied(g)

(I just invented the "satisfied" for the lack of a better notation)

Again, I prefer to say that {mi sisku da poi crino} adds an
object to the universe of discourse which satisfies {crino} and can map
to a number of physical objects, therefore creating the feeling of
a restricted universal quantification.

We do this in NatLangs as well: "I'm looking for a shirt.",
"I'm searching for something green.", "Ich suche eine Kuh.",
"Je cherche une vache.", ...

Just to point this out again: This is an analysis which I proposed in
my last mail and which I never directly read about anywhere. Therefore,
I'm still waiting for criticism and comments.

> This still supposes, of course, that there is something green in the UD,
> so the property sense is still better.

> Of course, spelling out the counterfactual stuff in such a way as to make
> the quantifier scope points clearer would be nice, too,
> but no one seems to like {tu'a}
> and it is a little iffy around the edges anyhow.

I don't really get, what you try to point out here.


v4hn

[John E Clifford]

Sorry, standard (in at least some groups I write in) logical notation: A for universal quantifier, S fo particular (L for salient, ? for interrogative, but thosedon't turn up here).  Quantifiers  take two wffs and a variable, AxFxGx is AllFsareGs, that universality restricted to the (non-null) extension of F.  or [x:Fx]Gx.  I suppose one could avoid the problem here by using  (x)(Green x => Seek I, x), but that doesn't really help.

I would be happy to have a better analysis of "seek", in particular, one that allowed for quantifiers to be placed properly without question, but I don't see it anywhere.  Much of the problem is in how we deal with intensional phrases.  Of the two usual approaches, having certain places specified as such in the lexicon or having all places transparent but some phrases labelled as intensional, Lojban has chosen a position in the middle.  All places are transparent, but some have recommended or required intensional phrase structures for filling.  Unfortunately, these cases don't cover all the intensional cases (and cover a number which are not intensional as well), so we are left with thing like thing {sisku} (which is not actually in Lojban, after all, but is popularly uses as though it were), where the transparent place yields unwanted results.

Ultimately, of course, what we want is a particular quantifier in the scope of the subjunctive, which is my informal summary of the role of {tu'a}.  So, for me, at least, {mi sisku tu'a da poi crino} means "I am looking for something green" with no hint that a particular one (or even one in the present UD) is required, since it expands to the more satisfying "I have a goal which would be satisfied just in case I were to have something green", with the quantifier tucked in the right place.  The standard explanation of {tu'a} gets close to this but gets bogged down in technicalities.

Your solution, as I understand it (if at all), is that {mi sisku da poi crino} is indeed transparent and the occurrence there of {da poi crino} may change the UD by adding an object to guarantee that the extension of {crino} is non-null.  If the extension of {crino} is already non-null, however, this object is to be identified with some already present object, which one depending on which one I actually find (more or less).  But that kind of anonymous object isn't allowed in the semantics game, nor does it help, since, as soon as its identity is revealed we fall back to the position of the external quantifier (which we never did really leave, if the slot was transparent), that I was really seeking this particular thing, not just any old thing at all.  Or, taking the broader view, I am really seeking every green thing individually.  Not what is wanted. 

 

---

From: v4hn <m...@v4hn.de>
To: loj...@googlegroups.com

Sent: Friday, January 11, 2013 4:42 AM
Subject: Re: [lojban] searching

[Ian Johnson]

Here's an idea I just had. I don't actually like it, but the fact that it works seems to say something about the issue. If {sisku} were defined as "x1 is searching among x3 and x1 would be satisfied if they found x2", then {mi sisku lo ckiku} does what was originally wanted while {mi sisku ro crino} does what {mi sisku lo ka crino} is defined to do. So this definition basically solves the problem (I think using {joi} you can specify that you would only be satisfied if you found several different sumti, in that (rather common) case. {.e} frustratingly doesn't work.)

This definition feels highly nonprimitive (though so does current {sisku}). In particular (in this regard unlike current {sisku}) it induces hidden quantifier/subjunctivity scope, which is rather important to what is actually meant. I'm pretty sure hiding such things is one of the major things we'd like to get away from with this language. 

Perhaps we should just derail this into a discussion of how best to handle subjunctivity? 

mi'e la latro'a mu'o

--
You received this message because you are subscribed to the Google Groups "lojban" group.
To post to this group, send email to loj...@googlegroups.com.

To unsubscribe from this group, send email to lojban+un...@googlegroups.com.

[John E Clifford]

Not sure how this helps, but there are two different issues here.  On the one hand, we need to deal with opaque phrases in the ordinary run of things; on the other hand we need to deal with contrary-to-fact situations as not ordinary run of things (though far more common than appears in most Lojban -- mainly because we are not sure how to do it).  Your suggestion is to reduce the first problem to the second (and then make it disappear back into the definition of words involved, so still available to surprise us).  But not all opaque cases are contrary-to-fact, we have the cases with {du'u} and {nu} and the like already (and regularly screw them up anyhow -- see raising).  The difficult cases are where we are not sure what abstraction is appropriate -- or even that one is, like thing {sisku} and {djica} and so on.  These very often are buried contrary-to-facts and for them we do have {tu'a}, stripped of its connection to (unspecified) buried abstractions and nebulous predicates, as a mark that the following term 1) cannot be moved or quantified out of its place (identified with things outside) and 2) at some point in an analysis will take its place in one or more alternate worlds which represent the working out of the predicate to which the term is attached as argument. 
The matter of contrary-to-fact or hypothetical sentences seems to involve just working out the rules on scope and the like for {da'i}.  I do not include the problems with truth conditions here, of course, since, so far as I can tell, no one has come up with a good answer to questions like "If Socrates were a 17th century Irish washerwoman, would Plato still have been gay?"

---

From: Ian Johnson <blindb...@gmail.com>

To: loj...@googlegroups.com
Sent: Friday, January 11, 2013 11:37 AM
Subject: Re: [lojban] searching

[la gleki]

On Friday, January 11, 2013 9:37:48 PM UTC+4, Latro wrote:

Here's an idea I just had. I don't actually like it, but the fact that it works seems to say something about the issue. If {sisku} were defined as "x1 is searching among x3 and x1 would be satisfied if they found x2", then {mi sisku lo ckiku} does what was originally wanted while {mi sisku ro crino} does what {mi sisku lo ka crino} is defined to do. So this definition basically solves the problem (I think using {joi} you can specify that you would only be satisfied if you found several different sumti, in that (rather common) case. {.e} frustratingly doesn't work.)

This definition feels highly nonprimitive (though so does current {sisku}). In particular (in this regard unlike current {sisku}) it induces hidden quantifier/subjunctivity scope, which is rather important to what is actually meant. I'm pretty sure hiding such things is one of the major things we'd like to get away from with this language. 

Perhaps we should just derail this into a discussion of how best to handle subjunctivity? 

mi'e la latro'a mu'o

For me sisku2 parallels jimpe2 and djuno2.

But djuno3 is often duplicated inside djuno2. So why doesnt sisku work like this 

x1 searches for x2 being x3 (nu).

{mi sisku lo ckiku lo ka zvati makau.}

"I'm searching for the keys, where they are located to be precise."

[Ian Johnson]

That doesn't really fix anything; with makau like that you are really just saying {mi sisku lo se zvati be lo ckiku}. That approach also doesn't get you access to indefinite searching, which is as has been noted occasionally necessary.

mi'e la latro'a mu'o

To view this discussion on the web visit https://groups.google.com/d/msg/lojban/-/r8er0cinPMYJ.

[Felipe Gonçalves Assis]

John has already made a very careful logical analysis of the topic,
so let me try something more concrete:

Saying that "I look for something green", in the sense that any green
thing will do, can be rendered as {mi sisku lo crino} for a convenient
definition of {sisku} is analogous to saying that "I know the result of
2+2" can be rendered as {mi djuno lo sumji be li re bei li re}, which
is actually expressing a relation between me and the number 4:
{mi djuno li vo}.

mu'o
mi'e .asiz.

[John E Clifford]

Hmmm! Yes.  Sorta.  The right referent but the wrong sense, light pointing to Venus shining just after sunset and saying "Lo, the Morning Star!"

---

From: Felipe Gonçalves Assis <felipe...@gmail.com>
To: loj...@googlegroups.com

Sent: Saturday, January 12, 2013 11:28 PM
Subject: Re: [lojban] searching

> lojban+unsub...@googlegroups.com.

> For more options, visit this group at
> http://groups.google.com/group/lojban?hl=en.
>
>
> --
> You received this message because you are subscribed to the Google Groups
> "lojban" group.
> To post to this group, send email to loj...@googlegroups.com.
> To unsubscribe from this group, send email to

> lojban+unsub...@googlegroups.com.

> For more options, visit this group at
> http://groups.google.com/group/lojban?hl=en.
>
>
> --
> You received this message because you are subscribed to the Google Groups
> "lojban" group.
> To post to this group, send email to loj...@googlegroups.com.
> To unsubscribe from this group, send email to

> lojban+unsub...@googlegroups.com.

> For more options, visit this group at
> http://groups.google.com/group/lojban?hl=en.

--
You received this message because you are subscribed to the Google Groups "lojban" group.
To post to this group, send email to loj...@googlegroups.com.

To unsubscribe from this group, send email to lojban+unsub...@googlegroups.com.

[Ian Johnson]

Assuming your "something" is indefinite (i.e. you'd be satisfied by whatever green thing), this is somewhat true, though I think xorlo makes it more slippery than your example. That is, {lo crino} could just be a blob of all the green things and could interact with sisku in such a way that you're satisfied by finding any of them.

mi'e la latro'a mu'o

[John E. Clifford]

Well, yes, {lo crino} might be the topmost node in the extension of {crino} and apply disjunctively as {sisku2}, but there still have to be green things for this to work, and there made not be -- hence the search, after all.

Sent from my iPad