x1 to the left

32 views
Skip to first unread message

Luke Bergen

unread,
Nov 12, 2010, 1:30:56 PM11/12/10
to loj...@googlegroups.com
It just fully occurred to me again how absurd this is.  What was the reasoning behind deciding that the x1 goes to the left of the selbri and if you want to say "mi prami do" with the selbri first you must say "prami fa mi do"?

It seems silly and arbitrary that if you just see "prami mi" you assume that the "mi" takes the x2.  It seems like the general rule should be "if you see a selbri, it's the selbri, if you see a sumti, make it the first sumti, if you already have a first sumti, make it the second sumti, etc..." until the whole bridi is understood but this general understanding of a bridi needs a special rule to handle "am I to the right or left of the selbri".  Why this bizarre rule?

mu'o mi'e cribe

Michael Turniansky

unread,
Nov 12, 2010, 1:59:36 PM11/12/10
to loj...@googlegroups.com
Because there are many, many times when you simply want to comment
about the existense of soemthing, or the state of things, and the
observative is a useful construct. So I can say "viska mi'o" without
having to say "da viska mi'o" Or "rokci" without having to say "lo
rokci cu farlu mi'o" Is it arbitrary? Sure, but arbitrary doesn't
equal bizzare. Back in the olden days of Loglan, something without
an argument to the left was considered to be a command, just like
English, (i.e., in modern lojban "broda" would have equalled "ko
broda"). It's just a design choice.

--gejyspa

> --
> 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.
>

Jorge Llambías

unread,
Nov 12, 2010, 2:06:52 PM11/12/10
to loj...@googlegroups.com
On Fri, Nov 12, 2010 at 3:59 PM, Michael Turniansky
<mturn...@gmail.com> wrote:
>  Back in the olden days of Loglan, something without
> an argument to the left was considered to be a command, just like
> English, (i.e., in modern lojban "broda" would have equalled "ko
> broda").  It's just a design choice.

Yes, I think that's the reason. Since commands were marked by the
absence of x1, for "dunda ta mi" to mean "give that to me" the
arguments to the right had to start from x2.

The other justification I heard is that you usually want ko'a in "lo
broda be ko'a", or in "da poi broda ko'a" to fall in the x2 place.

mu'o mi'e xorxes

Jonathan Jones

unread,
Nov 12, 2010, 3:15:06 PM11/12/10
to loj...@googlegroups.com
2010/11/12 Jorge Llambías <jjlla...@gmail.com>
Considering what gadri do, it seems perfectly sensible to me.


--
mu'o mi'e .aionys.

.i.a'o.e'e ko cmima le bende pe lo pilno be denpa bu .i doi.luk. mi patfu do zo'o
(Come to the Dot Side! Luke, I am your father. :D )

Frank

unread,
Nov 12, 2010, 3:35:20 PM11/12/10
to loj...@googlegroups.com
On 12 November 2010 13:59, Michael Turniansky <mturn...@gmail.com> wrote:
(i.e., in modern lojban "broda" would have equalled "ko
broda").  It's just a design choice.

              --gejyspa

However, in lojban, does not the absence of an argument to a selbri simply mean that argument is unspecified?  In such case, "prami mi" means "Someone unspecified Loves Me" and is not the command "Love Me".  In order to make it a command, if I recall correctly, One is required to use "ko", so as to make the statement, "You (imperative), Love Me."  Such an idiom ("absence equals unspecified") is seen in the former Google Wave lesson where the Student is instructed the phrase "zdani mi" means "something houses Me" or "I have a house" and is not told "You (imperative), house Me."  Is there actually an updated version of lojban, this "modern lojban", I have missed?  Or are You just saying, "If We were to start from scratch, We might make 'broda' equal 'ko broda'."?

Just wanting to make sure I have read materials correctly.

mi'e xuinkrbin.

Craig Daniel

unread,
Nov 12, 2010, 3:46:11 PM11/12/10
to loj...@googlegroups.com
2010/11/12 Jorge Llambías <jjlla...@gmail.com>:

>
> The other justification I heard is that you usually want ko'a in "lo
> broda be ko'a", or in "da poi broda ko'a" to fall in the x2 place.
>

That's the reason I like, even if it's not the historically correct
one for how it got that way.

But I personally think of selbri-last as the "default" order (with
lots of variation possible, either to make for less tangled syntax or
for pragmatically marked reasons) precisely because it makes x1 feel
less privileged.

On Fri, Nov 12, 2010 at 3:35 PM, Frank <frankdm...@gmail.com> wrote:
> On 12 November 2010 13:59, Michael Turniansky <mturn...@gmail.com> wrote:
>>

>> (i.e., in modern lojban "broda" would have equalled "ko
>> broda").  It's just a design choice.
>>

>>               --gejyspa
>
> However, in lojban, does not the absence of an argument to a selbri simply
> mean that argument is unspecified?  In such case, "prami mi" means "Someone
> unspecified Loves Me" and is not the command "Love Me".  In order to make it
> a command, if I recall correctly, One is required to use "ko", so as to make
> the statement, "You (imperative), Love Me."  Such an idiom ("absence equals
> unspecified") is seen in the former Google Wave lesson where the Student is
> instructed the phrase "zdani mi" means "something houses Me" or "I have a
> house" and is not told "You (imperative), house Me."  Is there actually an
> updated version of lojban, this "modern lojban", I have missed?  Or are You
> just saying, "If We were to start from scratch, We might make 'broda' equal
> 'ko broda'."?
> Just wanting to make sure I have read materials correctly.
> mi'e xuinkrbin.

Yup, you've got it right. Loglan did it differently, but Lojban works
exactly as you think.

- mi'e .kreig.daniyl.

Frank

unread,
Nov 12, 2010, 4:52:30 PM11/12/10
to loj...@googlegroups.com
On 12 November 2010 15:46, Craig Daniel <craigb...@gmail.com> wrote:

Yup, you've got it right. Loglan did it differently, but Lojban works
exactly as you think.

 - mi'e .kreig.daniyl.

Oh, good, because I am definitely liking this (relative to Me) new way of thinking.  ;-)

ki'e 

Lindar

unread,
Nov 12, 2010, 11:16:10 PM11/12/10
to lojban
My two cents:

It's a feature, and a rather useful one. Very often the x1 is omitted
to make it a more observation-like phrasing. Speak Lojbanically and
you'll see why we have it set up that way. =P

Luke Bergen

unread,
Nov 13, 2010, 1:46:38 AM11/13/10
to loj...@googlegroups.com
haha, I do speak lojbanically; as often as I can.  It's mostly the nit-picky theory-oriented side of me that's irked by the fact that all the sumti slots of a selbri are same except in this one way.  You use fa to move around to the slots or se to swap them around, all of this you do the same way for all of them.  They all live by the same rules.... except this one.

It's like "all dogs go to heaven... except for poodles... because it's convenient and therefore desirable".

It also seems somewhat biased against verb-object-subject and verb-subject-object types of languages.

Timo Paulssen

unread,
Nov 13, 2010, 3:53:13 AM11/13/10
to loj...@googlegroups.com
On 11/12/2010 09:35 PM, Frank wrote:
> On 12 November 2010 13:59, Michael Turniansky <mturn...@gmail.com>
> wrote:
>> (i.e., in modern lojban "broda" would have equalled "ko broda").
>> It's just a design choice.
>>
>> --gejyspa
>
>
> However, in lojban, does not the absence of an argument to a selbri
> simply mean that argument is unspecified? [�]
>
> mi'e xuinkrbin.

That's right. gejyspa just didn't want to use Loglan words, because many
people here probably don't know them. What he meant was: If Lojban now
was the way Loglan was back then, {broda} would equal what we now know
as {ko broda}.

mu'o mi'e timos

Ian Johnson

unread,
Nov 13, 2010, 10:28:29 AM11/13/10
to loj...@googlegroups.com
If you're using place structure, you're inevitably going to "discriminate" against some word order that is standard in some natlang. After all, there are 6 different ones: SVO, SOV, VSO, VOS, OVS, OSV. Lojban without FA or SE is SVO.

I will say, though, that VSO is a bit tempting; it takes the idea of selbri as predicates that take arguments and extends it to the actual way of speaking the language. The main issue that I would have is that we would be sacrificing the notion of a "bridi tail", which has some nice properties. One of them, albeit one that doesn't actually get used much (probably because most seljbobau are also selglibau) is the ability to say:
ko'a ko'e broda gi'e brode

A concrete example (this is used in L4B):
la nik. la kalifornias. klama gi'e stali

mu'o mi'e latros.

John E Clifford

unread,
Nov 13, 2010, 10:58:26 AM11/13/10
to loj...@googlegroups.com
Relevant to another thread, note that here a selbri is a predicate, not a relation.  Standard logic is VSO (though 'S' and 'O' introduce a non-logivcal distinction).  Shifting to SVO in logic is almost always a sign that something non-standard is about to happen (the difference between ~xF and x~F, for example--one of the reasons for SVO in Logjam).


From: Ian Johnson <blindb...@gmail.com>
To: loj...@googlegroups.com
Sent: Sat, November 13, 2010 9:28:29 AM
Subject: Re: [lojban] Re: x1 to the left

Pierre Abbat

unread,
Nov 13, 2010, 2:07:22 PM11/13/10
to loj...@googlegroups.com
On Saturday 13 November 2010 10:28:29 Ian Johnson wrote:
> If you're using place structure, you're inevitably going to "discriminate"
> against some word order that is standard in some natlang. After all, there
> are 6 different ones: SVO, SOV, VSO, VOS, OVS, OSV. Lojban without FA or SE
> is SVO.

Lojban can do SVO and SOV equally well. Unlike natlangs, however, it doesn't
have a small number of places for arguments that are usually assigned to
subject, direct object, and indirect object; rather, it has an indefinite
sequence of places. x3 of dunda corresponds to the indirect object, but there
is no way x3 of mlafi'e (the side that has two eyes) could be called an
indirect object.

Pierre
--
When a barnacle settles down, its brain disintegrates.
Já não percebe nada, já não percebe nada.

Michael Turniansky

unread,
Nov 14, 2010, 9:19:27 AM11/14/10
to loj...@googlegroups.com
On Sat, Nov 13, 2010 at 3:53 AM, Timo Paulssen
<timo...@perpetuum-immobile.de> wrote:
> On 11/12/2010 09:35 PM, Frank wrote:
>> On 12 November 2010 13:59, Michael Turniansky <mturn...@gmail.com>
>> wrote:
>>> (i.e., in modern lojban "broda" would have equalled "ko broda").
>>> It's just a design choice.
>>>
>>> --gejyspa
>>
>>
>> However, in lojban, does not the absence of an argument to a selbri
>> simply mean that argument is unspecified? […]

>>
>> mi'e xuinkrbin.
>
> That's right. gejyspa just didn't want to use Loglan words, because many
> people here probably don't know them. What he meant was: If Lojban now
> was the way Loglan was back then, {broda} would equal what we now know
> as {ko broda}.
>
> mu'o mi'e timos
>
Yes, hence the "would have" (subjunctive (approx. = va'o da'i lo
nu...)) On the other hand, timos, there are plenty of Loglan words
lojbanists know -- bisli, rismi, vamtu, blanu (although blanu had an
x2 place back in Loglandia (x1 is bluer than x2))
--gejyspa

Jorge Llambías

unread,
Nov 14, 2010, 9:31:46 AM11/14/10
to loj...@googlegroups.com
On Sun, Nov 14, 2010 at 11:19 AM, Michael Turniansky
<mturn...@gmail.com> wrote:
>   On the other hand, timos, there are plenty of Loglan words
> lojbanists know -- bisli, rismi, vamtu, blanu (although blanu had an
> x2 place back in Loglandia  (x1 is bluer than x2))

And there are even some words they might think they recognize but are
false friends, like "cidja" (means "awake" in Loglan) or "barda"
(means "rewards" in Loglan).

Reply all
Reply to author
Forward
0 new messages