--
You received this message because you are subscribed to the Google Groups "Lojban Beginners" group.
To view this discussion on the web visit https://groups.google.com/d/msg/lojban-beginners/-/tNjlAG8IEs8J.
To post to this group, send email to lojban-b...@googlegroups.com.
To unsubscribe from this group, send email to lojban-beginne...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/lojban-beginners?hl=en.
I think you mean "do cusku lo sedu'u mi prami do". The second "do"refers to the same person as the first "do". It would be otherwise if you had
said "do cusku lu mi prami do li'u".
Pierre
--
gau do li'i co'e kei do
Definition of {du'u}:
x1 is predication [bridi] expressed in sentence x2
So {se} works not directly on {du'u}, but on {du'u broda}, wich means
'x1 is a fact of broda expressed in sentence x2'. Then {lo du'u broda}
means 'a fact of broda' and {lo se du'u} means 'a sentence expressing
a fact of broda'.
mu'o mi'e ianek
On 7 Sie, 06:02, Pierre Abbat <p...@phma.optus.nu> wrote:
> On Monday 06 August 2012 23:51:58 Luke Bergen wrote:
>
> > Is that true? If it were "lo du'u mi prami do" it would be true for sure.
> > But isn't "sedu'u" the text which expresses the bridi x1?
>
> > sedu'u has always kind of confused me. It's not a bridi, it's a "text
> > expressing [bridi] which is x2", but not the kind of "text" that you get
> > with lu li'u?
>
> "do cusku lo sedu'u mi prami do" is equivalent to "do cusku lu do prami mi
> li'u" (except that in the first case it could have been said in Greek or
> Ithkuil or anything else).
>
> Pierre
>
> --
> Jews use a lunisolar calendar; Muslims use a solely lunar calendar.
--
You received this message because you are subscribed to the Google Groups "Lojban Beginners" group.
To post to this group, send email to lojban-b...@googlegroups.com.
To unsubscribe from this group, send email to lojban-beginne...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/lojban-beginners?hl=en.
"cmavo clusters" in jbovlaste are somewhat messy, from what I've seen.
http://camxes.lojban.org/camxes/?text=se+du%27u+broda
You basically confirmed what I was saying, but you formally disagreed
with me. I don't understand. {se} acts on selbri, BAI, JA and some
other things, but not on NU. It acts on selbri built with NU, just
like in your examples.
On 7 Sie, 09:47, Jonathan Jones <eyeo...@gmail.com> wrote:
> On Tue, Aug 7, 2012 at 1:15 AM, ianek <jane...@gmail.com> wrote:It's easier pe'i to see it as switching x1 and x2 of {du'u broda},
> > "cmavo clusters" in jbovlaste are somewhat messy, from what I've seen.
> >http://camxes.lojban.org/camxes/?text=se+du%27u+broda
>
> The {se} still acts on {du'u}, not {du'u broda}.
>
> Whatever comes after a NU, until the {kei} (elided or not), is "eaten" by
> the NU.
>
> So {[zo'e] du'u broda [kei] [zo'e]} means "x1 is predication "broda"
> expressed in sentence x2".
>
> {se} merely switches the x1 and x2 of {du'u}.
because places are primarily property of selbri, not abstraction
markers.
> So, for example:Yes, and how it proves me wrong? I know all that, of course. {du'u mi
>
> du'u mi nelci lo blanu kei zoi gy. I like Blue. .gy
>
> and
>
> .i zoi gy. I like Blue. .gy se du'u mi nelci lo blanu
>
> are equivalent.
nelci lo blanu kei} is a selbri with two places and can be converted
with se. Why do we need to assume that SE act on NU if such assumption
is not needed to explain the grammar?
I love how you still treat me as a complete co'a cilre.
On Tue, Aug 7, 2012 at 4:50 AM, ianek <jan...@gmail.com> wrote:
You basically confirmed what I was saying, but you formally disagreed
with me. I don't understand. {se} acts on selbri, BAI, JA and some
other things, but not on NU. It acts on selbri built with NU, just
like in your examples.
{se} acts on /anything/ with an x2. {du'u} is a NU which has an x2.
On 7 Sie, 09:47, Jonathan Jones <eyeo...@gmail.com> wrote:
> On Tue, Aug 7, 2012 at 1:15 AM, ianek <jane...@gmail.com> wrote:
> > "cmavo clusters" in jbovlaste are somewhat messy, from what I've seen.
> >http://camxes.lojban.org/camxes/?text=se+du%27u+broda
>
> The {se} still acts on {du'u}, not {du'u broda}.
>It's easier pe'i to see it as switching x1 and x2 of {du'u broda},
> Whatever comes after a NU, until the {kei} (elided or not), is "eaten" by
> the NU.
>
> So {[zo'e] du'u broda [kei] [zo'e]} means "x1 is predication "broda"
> expressed in sentence x2".
>
> {se} merely switches the x1 and x2 of {du'u}.
because places are primarily property of selbri, not abstraction
markers.
Abstraction markers /are/ selbri. That's why they have places.
> So, for example:Yes, and how it proves me wrong? I know all that, of course. {du'u mi
>
> du'u mi nelci lo blanu kei zoi gy. I like Blue. .gy
>
> and
>
> .i zoi gy. I like Blue. .gy se du'u mi nelci lo blanu
>
> are equivalent.
nelci lo blanu kei} is a selbri with two places and can be converted
with se. Why do we need to assume that SE act on NU if such assumption
is not needed to explain the grammar?
Because it's not an assumption. It's a fact.
I love how you still treat me as a complete co'a cilre.
No, I treat you like some who apparently needs to read the relevant section of the official text on our language.
-- .i pau mi me ma .i pa mai ko mi jungau la'e di'u .i ba bo mi va'o lo nu nelci lo nu me ma kau cu barkla .i va'o lo nu na nelci cu denpa ti lo nu mi drata
Am 07.08.2012 20:47, schrieb Jonathan Jones:
No. Despite the wording the CLL uses, a NU itself has no places at all.On Tue, Aug 7, 2012 at 4:50 AM, ianek <jan...@gmail.com> wrote:
You basically confirmed what I was saying, but you formally disagreed
with me. I don't understand. {se} acts on selbri, BAI, JA and some
other things, but not on NU. It acts on selbri built with NU, just
like in your examples.
{se} acts on /anything/ with an x2. {du'u} is a NU which has an x2.
--
You received this message because you are subscribed to the Google Groups "Lojban Beginners" group.
To post to this group, send email to lojban-b...@googlegroups.com.
To unsubscribe from this group, send email to lojban-beginne...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/lojban-beginners?hl=en.
--
You received this message because you are subscribed to the Google Groups "Lojban Beginners" group.
To post to this group, send email to lojban-b...@googlegroups.com.
To unsubscribe from this group, send email to lojban-beginne...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/lojban-beginners?hl=en.
On Tue, Aug 7, 2012 at 3:51 PM, Jonathan Jones <eye...@gmail.com> wrote:Jonathan, you are obviously a more knowledgeable person on matters of
> Really? Then why do they ALL have place tags in their definition?
>
> du'u abstractor: predication/bridi abstractor; x1 is predication [bridi]
> expressed in sentence x2.
Lojban than myself; so pardon me whether I'm asking a stupid question.
I want to ask why those definitions contain *[bridi]* in their
definitions.
By themselves, I assume they don't. I must assume thus
that the definition refers to a complete abstraction and not the
abstractor by itself.
How am I wrong? Does any abstractor form an
abstraction all by itself?
If so, what {lo se du'u kei} would mean?
Does it mean {lo se du'u co'e kei}?
Or is it illegal?
Please, explain.
mu'o mi'e betsemes
--
You received this message because you are subscribed to the Google Groups "Lojban Beginners" group.
To post to this group, send email to lojban-b...@googlegroups.com.
To unsubscribe from this group, send email to lojban-beginne...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/lojban-beginners?hl=en.