mi kakne lo bajra

[Oren]

Can someone clarify why:

sentence A:

{ mi kakne lo bajra }

is not as good as

sentence B:

{ mi kakne lo nu bajra }

Assuming the rationale is that "nu" is necessary to mark an event/state, it seems to me that since (event/state) is inherent in the x2 argument of kakne, that adding the 'nu' would be superfluous.

Assuming the argument is that it becomes nonsensical to say sentence A, i.e. "You can't say I can runner," I would think that argument is based on English concepts of proper grammar, and that there should be no a priori reason something like "I can runner" is "nonsensical."

Or is there some other veciski?

[Michael Turniansky]

Since the x2 of kakne is an event, the first sentence says you are
capable of some "running event". That is, some event that runs
(whatever that maybe intended to mean), not an event OF running. For
a less confusing contrast:
a)mi kakne lo drata

b)mi kakne lo nu drata

The first says "I am capable of other things". The second second
says "I am capable of being different"

--gejyspa

[Jameson Orndorff]

{.i mi kakne lo bajra}, if you translate it almost literally, means "I
am capable of a runner." Even if {kakne} didn't require an abstraction
specifically, I still think kakne2 wants a bridi for an answer.
Otherwise, that's why we have {tu'a}.

> --
> 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.
>
>

[Michael Turniansky]

On Tue, Oct 26, 2010 at 2:06 PM, Michael Turniansky
<mturn...@gmail.com> wrote:

To clarify a bit (I was heading out the door when I was wrote the
above) -- kakne means "x1 can do an event/state x2 under conditions
x3". X2 IS an event or state, regardless of what it might be. If x2
starts with "lo nu", "lo za'i", etc. then it's essentially superfluous
(other than to clarify if it's an event or a state). But if it
doesn't such as "lo bajra", then it is an event or a state AS WELL AS
being a runner.

I also might not have been very clear what the difference of my a)
and b) above were. So let me give it some context:

a) mi bajra .i mi kakne lo drata (I am capable of a event which is different)

b) mi bajra .i mi kakne lo nu drata. (I am capable of the event of
me being different)

a) I am running. I can do something else (besides running).
b) I am running. I can (also) be something other (than what I am,
presumably, but not necessarily referring to runningness).

--gejyspa

[Oren]

Thanks for the replies and explanations! The lo drata / lo nu drata
comparison helps a great deal.

But I don't yet fully comprehend "X2 IS an event or state, regardless

of what it might be.  If x2 starts with "lo nu", "lo za'i", etc. then
it's essentially superfluous
(other than to clarify if it's an event or a state).  But if it
doesn't such as "lo bajra", then it is an event or a state AS WELL AS
being a runner."

...Part of me feels like, at least when we look at the English
glosses, that the "lo drata/ lo nu drata" example has a readily
apparent semantic difference. To help me understand the generality of
this contrast, what precisely would be semantically different in the
original two cases. It seems to me that the two should be equivalent:

Lojban: { mi kakne lo bajra }
Parse: (x1: I) am capable of (x2 [FORCE-EVENT/STATE: "being a"] runner")
Gloss: "I'm capable of a running event"
English: (I can run)

Lojban: { mi kakne lo nu bajra }
Parse: (x1 I) am capable of (x2 [EXPLICIT-EVENT/STATE: "being a"] runner")
Gloss: "I'm capable of the event of (me?*) running"
English: (I can run)

ki'e .i co'o mi'e korbi

[Pierre Abbat]

On Wednesday 27 October 2010 08:54:22 Oren wrote:
> Lojban: { mi kakne lo bajra }
> Parse: (x1: I) am capable of (x2 [FORCE-EVENT/STATE: "being a"] runner")
> Gloss: "I'm capable of a running event"
> English: (I can run)

A running event, here, is not an event of running. It is an event that runs.
How an event can run is not clear, so what this sentence means is not clear.

Another example of forcing: The preposition "ti'u" and the tense marker "ca"
force their object to be a time. (There is a semantic difference between them
which isn't relevant to forcing.) For example:
le nunpenmi be ti'u li vo pe le vanci cu fasnu ku'i ca le solnuncanci
The meeting set for four in the afternoon happened instead at sunset.
"ca le solnuncanci" is obvious: a sunset is an event which occurs at a
time. "ti'u li vo pe le vanci" is not quite so obvious: numbers neither are
times nor belong to evenings. One has to refer to a clock to make sense of
this. One can equally grammatically say "ca le skami", which forces the
computer to be a time just as "ti'u li vo" forces four to be a time. But what
time is a computer?

Pierre
--
lo ponse be lo mruli po'o cu ga'ezga roda lo ka dinko

[Michael Turniansky]

Part of your problem in understanding might be that, in my haste, I
dropped a word. I apologize. I meant to type ""x1 can do an
event/state x2 under conditions
x3". X2 IS an event or state, regardless of what *else* it might be"
(left out the "else"). So as Pierre says, " mi kakne lo bajra" means
that the _event_ is a running/is a runner. Maybe that means,
metaphorically, that you are capable of carrying out deeds that go by
quickly. But it can only be understood on a metaphorical level, and
not literally, since events are not normally things that run.

--gejyspa

[Luke Bergen]

I think part of your confusion is coming from your glosses.  I think you're incorrectly inserting the "being a" in your gloss of {mi kakne lo bajra}.  {mi kakne lo bajra} does not say "I am capable of [being a] runner".  It's just {I am capable of doing the 'runner' event}  whatever event "runner" might be.  You are not "being" anything when you say {mi kakne lo bajra}

On Wed, Oct 27, 2010 at 8:54 AM, Oren <get....@gmail.com> wrote:

[Adam Lopresto]

To me it sounds like you don't quite understand. It isn't that kakne2
*forces* an event, it's that nothing else makes any sense in that
place. In the same way, only things with legs can fit into bajra1. So
if you assert *{mi kakne lo bajra} or {lo se kakne cu bajra}, you're
talking about something that can fill kakne2 and bajra1 at the same
time, that is, something that is an event someone is capable of, that
*also* runs on legs on some surface. Since events or states don't have
legs, you've ended up with semantic nonsense, though it's fine
grammatically.

It's not that we add {nu} into {lo bajra} to turn {lo bajra} into an
event. Instead, build it from the inside out. {bajra} is a selbri
meaning "x1 runs on surface x2 with limbs x3 and gait x4". {nu ...
kei} wraps an entire bridi and converts it into a selbri, so {nu
[zo'e] bajra [zo'e] [zo'e] [zo'e] [kei]} is a selbri meaning "x1 is an
event of someone running on some surface with some gait". Then we
convert that to a sumti by surrounding it with {lo ... ku}, to create
a sumti that fills the x1 of that, so "some event or events of someone
running on some surface with some gait". And that event (and not a
runner) is the sort of thing that can fill kakne2.

[.alyn.post.]

Thank you Adam, this is a wonderful description.

-Alan

--
.i ko djuno fi le do sevzi

[Lindar]

I would just like to contribute a bit here.

There is no implicit event-ness of the x2 of kakne, it's just that
it's required to make sense.
{.i mi kakne lo bajra} literally means something like "I am capable of
a runner.".
No sense is made.

So {lo bajra} is a runner. It is something which runs.
{lo nu bajra} is running (or "an event of being a runner"). It is the
actual event itself.
I am capable of an event of being a runner. {.i mi kakne lo nu bajra}

So... your meaning is strange.
Come bug us on IRC if you want to talk about it in real-time.

[Krzysztof Sobolewski]

Dnia czwartek, 28 października 2010 o 06:03:26 Lindar napisał(a):
> There is no implicit event-ness of the x2 of kakne, it's just that
> it's required to make sense.
> {.i mi kakne lo bajra} literally means something like "I am capable of
> a runner.".
> No sense is made.

I would not be surprised if hypothetical future native speakers of lojban would find this shortcut very convenient, though...
--
Ecce Jezuch
"All of us get lost in the darkness
Dreamers learn to steer by the stars
All of us do time in the gutter
Dreamers turn to look at the cars" - N. Peart

[Lindar]

> I would not be surprised if hypothetical future native speakers of lojban would find this shortcut very convenient, though...

Is not shortcut, is wrong.
Not think is good.

>_>

It's not a shortcut, it's just wrong. It means "I'm capable of a
runner." and that's always what it will mean. Nobody that -actually-
speaks Lojban would agree with your sentiments. lo bajra will always
be a runner, and lo nu bajra will always be an event of being a runner
(or "running"), and the day that it stops working like that is the day
that I give up on Lojban forever. It would introduce vast amounts of
ambiguity to have implicit abstractors, and has always been rejected.
That's called "sumti raising" and is probably worse than malgli.

How would it be a shortcut, anyway? It's one phoneme.

(Sorry for getting huffy. I did this when I was new as well. Just... -
don't- do it. We do things the way they're done for a very good
reason, and we've been doing it for at least a score longer than you.
Rather than suggest we change something, learn why we do it.)

[Oren]

Alright, I understand now. My confusion was motivated in part by my
glosses seeming indistinguishable. The demonstration that {mi kakne lo
bajra } implies {lo se kakne cu bajra } cleared this up for me a great
deal!

However, I got a bone to pick now! As someone translating the gismu
list into Chinese, I wish I had known that the
(parenthetical/classifiers) before sumti were (sometimes?)
prerequisites for "sensical-ness," and not just helpful guides like
the [square/brackets/of/synonyms] before some sumti are.

I'm specifically thinking of the linked gismu list found on jbotcan
[http://jbotcan.org/gismulinked.html], where these parenthetical
expressions can range from as precise and seemingly "highly
recommended" as (jo'u) or (ka), to as English and seemingly only for
glossing purposes as (person/object), (event/state) or even (agent).

I may just be missing something. Are there in fact explicit
recommended markers for *all* (parenthetical/classifiers) on the list
I'm referencing? That is, it seems here that "... (event/state) x2
..." means something like "... (use nu or za'i) x2 ..." ...do such
cmavo-prescriptions exist for (event/property/interval/idea) too? And
all the rest?

I suspect that for things like (object/agent) we may be relying on
gray-area "common sense" distinctions between groups of gismu (i.e.
"Well, that makes no sense because normally tables aren't agents!"),
and not anything formalized-- in which case, I think we should improve
this document to make that clearer, even if it just means explicitly
using cmavo for instances where cmavo are recommended, and leaving the
other (gloss/classifiers) alone.

Although if we all had more time, I think it would be valuable to come
up with lists of sensical "object/agent/etc" tags for gismu.

co'o mi'e korbi

> --
> 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.
>
>

--
Oren Robinson
(315) 569-2888
102 Morrison Ave
Somerville, MA 02144

[Michael Turniansky]

Well, with the caveat that (as I showed before with "mi kakne lo
drata") there is pretty much NEVER a place that OBLIGATES a use of a
particular cmavo, as the long as the sumti is in that class (simple
example, they pretty much can all take "ri" or "la'e di'u" where the
reference is of the proper class), I will start you off with
often-associated cmavo of sumti place descriptors

event nu
state za'i
idea si'o
property/quality ka
quanitifier li
set lo'i/le'i/la'i (although in the xorlo era, lo/le is probably
acceptable) or ....ce...
quantity ni
process pu'u
--gejyspa

[Adam Lopresto]

Again, though, it's *not* about the cmavo, it's about the type of
sumti. {lo bajra} is something that runs. {lo nu bajra} is an event.
But {lo nandu} is *also* an event. So there's nothing wrong with, say,

mi kakne lo nandu be do [be'o] [ku] [vau]
I can do things that are hard for you.

Now, the most common way to get an event sumti is to use {lo nu}, but
all that's relevant is the semantic type of the sumti. With that in
mind, (event) and (agent) are pretty much the same, in that they both
impose semantic limitations or guidance on what sort of thing makes
sense to fill that role.

[tijlan]

On 26 October 2010 20:43, Michael Turniansky <mturn...@gmail.com> wrote:
>
>  I also might not have been very clear what the difference of my a)
> and b) above were.   So let me give it some context:
>
> a)  mi bajra .i mi kakne lo drata (I am capable of a event which is different)
>
> b)  mi bajra .i mi kakne lo nu drata.  (I am capable of the event of
> me being different)
>
>  a) I am running.  I can do something else (besides running).
>  b) I am running.  I can (also) be something other (than what I am,
> presumably, but not necessarily referring to runningness).

Is the x1 of {drata} in {mi kakne lo nu drata} necessarily the x1 of
the main selbri i.e. {mi}?

[Michael Turniansky]

Necessarily? No. But it's commonly understood to mean that when
it's elided, yes.
--gjeyspa

[tijlan]

On 28 October 2010 10:34, Lindar <lindar...@yahoo.com> wrote:
>> I would not be surprised if hypothetical future native speakers of lojban would find this shortcut very convenient, though...
>
> Is not shortcut, is wrong.
> Not think is good.
>
>>_>
>
> It's not a shortcut, it's just wrong. It means "I'm capable of a
> runner." and that's always what it will mean. Nobody that -actually-
> speaks Lojban would agree with your sentiments. lo bajra will always
> be a runner, and lo nu bajra will always be an event of being a runner
> (or "running"), and the day that it stops working like that is the day
> that I give up on Lojban forever. It would introduce vast amounts of
> ambiguity to have implicit abstractors, and has always been rejected.

How exactly is "a runner" different from "an event of being a runner"?
If X is a runner, doesn't X represent an event of being a runner?

[Michael Turniansky]

No. X is a person (or animal, or whatever). An "event of running"
is an action. X can perform the action of running ("da zukte lo nu
[da] bajra") or can cause it ("da gasnu lo nu lo xirma cu bajra"), but
he can't BE it.
--gejyspa

[Krzysztof Sobolewski]

Dnia czwartek, 28 października 2010 o 11:34:09 Lindar napisał(a):

> It's not a shortcut, it's just wrong. It means "I'm capable of a
> runner." and that's always what it will mean.

Of course, but there are things like idioms, which are nonsensical literally, but do have a meaning. I'm not proposing to make it legal, don't worry, although I am slightly encouraged by cimjvo, but my main point is: how do you know what rules of the language will be inferred by a child's brain in the process of becoming a native speaker? I have no idea, I'm a liguistic noob and I don't have a clue what I'm talking about, so I'll stop here, but I sure would like to see that :)

P.S.: I too am not thrilled by sumti raising.
--
Ecce Jezuch
"Life feeds on life feeds on life feeds on life feeds on life...
This is necessary." - M. J. Keenan

[.alyn.post.]

I think "Dreshcher and the toaster" may be worth understanding:

http://www.catb.org/jargon/html/koans.html#id3141308

Drescher and the toaster

A disciple of another sect once came to Drescher as he was eating
his morning meal.

"I would like to give you this personality test", said the outsider,
"because I want you to be happy."

Drescher took the paper that was offered him and put it into the
toaster, saying: "I wish the toaster to be happy, too."

"I would like to give you this personality test because taking it
will result in you being happy." is one fix for one possible
interpretation of this koan. I know there are others! I do think
his koan demonstrates the problem domain.

Is happiness a property of the test?
Is toaster happiness the same as tester happiness?

-Alan

[Adam Lopresto]

> How exactly is "a runner" different from "an event of being a runner"?
> If X is a runner, doesn't X represent an event of being a runner?

.i tu cipni .i tu vasxu .i tu vofli .i ku'i tu ba'e na nu vasxu .i tu
na nu vofli
.i mi kakne lo nu vofli .i ku'i mi ba'e na kakne lo nu vofli .i na'i
mi kakne lo cipni no'u lo vasxu zi'e no'u lo vofli

That's a bird. It breathes. It flies. But it is not an event of
breathing, or an event of flying.
I am capable of an event of breathing. I am *not* capable of an event
of flying. And it's just nonsense to say I'm capable of a bird (which
is a breather, and a flyer).

[Ian Johnson]

How is that a shortcut? You do something confusing and save 1 tiny syllable. Possibly 2 if you need to terminate the abstractor.

mu'o mi'e latros.

2010/10/28 Krzysztof Sobolewski <jez...@interia.pl>

[Remo Dentato]

On Thursday, October 28, 2010, .alyn.post.

<alyn...@lodockikumazvati.org> wrote:
>
> Is happiness a property of the test?
> Is toaster happiness the same as tester happiness?

In the world described by the koan, I would say that the test has the
property of making someone happy (through an unspecified process) and
the toaster has (or is considered to have) the property of possibly
been happy.

[Pierre Abbat]

On Thursday 28 October 2010 13:13:49 .alyn.post. wrote:
> Is happiness a property of the test?
> Is toaster happiness the same as tester happiness?

I think the tester meant that he could tell, from the results of the test,
what would make Drescher happy. I'd have to know more about Drescher to make
sense of his action.

Pierre
--
Don't buy a French car in Holland. It may be a citroen.

[Jorge Llambías]

On Thu, Oct 28, 2010 at 2:13 PM, .alyn.post.
<alyn...@lodockikumazvati.org> wrote:
>
>  "I would like to give you this personality test", said the outsider,
>  "because I want you to be happy."
>
>  Drescher took the paper that was offered him and put it into the
>  toaster, saying: "I wish the toaster to be happy, too."

Drescher seems to think that "give test to someone/something" is
equivalent to "put test into someone/something". The joke wouldn't
work in Lojban very well, because I don't think you would say "dunda
lo cipra".

mu'o mi'e xorxes

[Oren]

It doesn't seem right to me that there are all of these implicit types
associated with certain gismu, without my knowledge. I feel like,
especially for learners of lojban, these virtual "parts-of-speech"
should be explicitly laid out, so that people don't have to question
whether "running" is an event or an action or what have you.

So... If anyone wants to help, I'm going to do this and integrate it
into the Chinese translation for the gismu list I'm doing.

https://spreadsheets.google.com/ccc?key=0AoZKIeNnTBe2dFAtNHNmWTlyeFR1TzIzaUZwMXNWeEE&hl=en&authkey=CNCr2d0B

[Ian Johnson]

They're not usually implicit. Most of the definitions in the gimste have things like (event/state) or (ka) or (du'u) labeled in them when it doesn't make sense for them to be something concrete. Even then, it still doesn't make sense for a {se kakne} to be concrete; how can you be capable of doing a concrete thing?

mu'o mi'e latros.

[Lindar]

On Oct 29, 10:45 am, Oren <get.o...@gmail.com> wrote:
> It doesn't seem right to me that there are all of these implicit types
> associated with certain gismu, without my knowledge.

1. It's written out in the definition very plainly.
djuno - x1 knows fact(s) x2 (du'u) about subject x3 by epistemology
x4.

2. It's almost common sense. We're not trying to sneak something past
you.
{.i mi djuno lo bajra} means something like "I know that a runner."
{.i mi djuno lo du'u do bajra} means something like "I know that you
run."

One of these doesn't make sense.

Some of them (like {fanza} for example) may not be too obvious due to
natural language bias, which is why we -do- have them explicitly laid
out, but common sense should still be applied. You can't want an
apple, you can want to -eat- an apple. You can't be annoyed by
somebody, you can be annoyed by something that somebody does.

> I feel like,
> especially for learners of lojban, these virtual "parts-of-speech"
> should be explicitly laid out, so that people don't have to question
> whether "running" is an event or an action or what have you.

It is. We should probably update the language in the definitions to
explicitly say (ka/nu/za'i) instead of (property/event/state), but
they already do have these explicitly stated in the definitions.

> So... If anyone wants to help, I'm going to do this and integrate it
> into the Chinese translation for the gismu list I'm doing.

I don't speak a bit of Chinese.

I do, however, appreciate your input and concern. I will document this
in the appropriate BPFK section and it will be addressed when all
other issues are as well. The idea behind these is explicitly laid out
very early on in the teaching model that we're trying to perfect, so
it will definitely be made more clear in future documents.

[Jorge Llambías]

On Fri, Oct 29, 2010 at 6:54 PM, Lindar <lindar...@yahoo.com> wrote:
>
> You can't want an
> apple, you can want to -eat- an apple.

Fortunately nobody pays any attention to the equally silly note in the
definition of "dunda" that you can't give an apple, you can only give
possession of an apple, or in "vecnu" that you can't sell an apple,
you can only sell the possession of an apple, or that you can't offer
("friti") an apple, and so on in several other gismu. People only seem
to fixate on poor "djica".

And eating it is not the only reason you may want an apple for. What
you want it for is what goes in the x3 of djica, the x2 of djica is
for what you want.

[Oren]

@Ian and Lindar: Let me rephrase this you don't seem to hear what I'm saying.

"1. It's written out in the definition very plainly.
djuno - x1 knows fact(s) x2 (du'u) about subject x3 by epistemology
x4."

I don't just mean there's an issue with how the sumti are labelled in
the gismu list. Those are trivial to modify. I mean there's no way of
knowing what gismu have those ka's and nu's built into them, and which
need them explicitly added in certain sumti positions.

For example, take a look at nandu and bajra.

{ mi kakne lo bajra } is nonsensical
{ mi kakne lo nu bajra } is sensical

{ mi kakne lo nandu } is sensical
{ mi kakne lo nu nandu } (I assume) is nonsensical or overspecified.

That is, while its true that "There are no conventional
parts-of-speech distinctions like adjectives or nouns in Lojban,"
there are still undeniable semantic roles that we expect -- and reject
-- from gismu when allocated to sumti placement.

I think it's great that we all know from the definition that { se
kakne } is an event/state (so it should have a nu), but I don't like
that there are unwritten rules for which gismu have event/state built
into them. For example, I wouldn't have anticipated the bajra/nandu
discrepancy shown above. I wish for something like:

bajra - is a process (pu'u), is a state (za'i)
nandu - is an event (nu)

See what I mean?

2010/10/29 Jorge Llambías <jjlla...@gmail.com>:

> --
> 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.
>
>

--

[Ian Johnson]

{lo bajra} runs. Why would the basic word involved in defining running be abstract, when there is a concrete concept of running that also warrants description? Similarly, what is {lo nandu}? Not a concrete thing, an abstraction; the notion "I am difficult" doesn't make sense when taken literally, which is why we have {jai}.

mu'o mi'e latros.

[Lindar]

Well, nandu should have the x1 place explicitly marked. It's an
abstraction!

For a dumb example...

.i lo djuno ku djuno lo se djuno

vs.

.i lo djuno ku djuno lo du'u se djuno kei ku

Example 1: Due to the fact that the x2 of djuno is a du'u, and the x1
of se djuno is a du'u, there's no abstraction necessary.

Example 2: This means something like "The knower knows that
(something) is a fact.", whereas the first example is "The knower
knows a fact.".

When abstracted places match up, no abstraction is necessary for the
example 1 meaning. Adding an abstraction changes the meaning.

That {nandu} isn't marked explicitly as having an abstracted place is
probably just an error.

xorxes: Sumti raising is sumti raising. There's no nitpicking or
arguing semantics about it. Wanting an apple for the purpose of eating
it is still sumti raising, because it's adding an implied concept of -
having-. That's what sumti raising is. =/

[Michael Turniansky]

On Sat, Oct 30, 2010 at 7:59 AM, Lindar <lindar...@yahoo.com> wrote:
> Well, nandu should have the x1 place explicitly marked. It's an
> abstraction!
>

(It does in the Spanish version written by xorxes at least 7 years ago...)

--gejyspa

[Adam Lopresto]

On Fri, Oct 29, 2010 at 11:57 PM, Oren <get....@gmail.com> wrote:
> @Ian and Lindar: Let me rephrase this you don't seem to hear what I'm saying.
>
> "1. It's written out in the definition very plainly.
> djuno - x1 knows fact(s) x2 (du'u) about subject x3 by epistemology
> x4."
>
> I don't just mean there's an issue with how the sumti are labelled in
> the gismu list. Those are trivial to modify. I mean there's no way of
> knowing what gismu have those ka's and nu's built into them, and which
> need them explicitly added in certain sumti positions.

No way other than how they're labeled in the gismu list, or by reading
and understanding them. But again, I think the fundamental problem is
that it isn't that some have a {ka} or {nu} *built in*, it's that
they're of different types.

> For example, take a look at nandu and bajra.
>
> { mi kakne lo bajra } is nonsensical
> { mi kakne lo nu bajra } is sensical

Right. bajra1 is something that runs, {nu bajra kei}1 is an event.
kakne2 is an event, not something that runs.

> { mi kakne lo nandu } is sensical
> { mi kakne lo nu nandu } (I assume) is nonsensical or overspecified.

Yes. The definition of {nandu} probably should be rephrased to make it
more clear that nandu1 is an event (the thing you said you didn't have
trouble with). If we did define it as

x1 (event) is difficult/hard/challenging for x2 (agent) under
conditions x3 (event/state)

then would it make more sense to you?

But again, the distinction between events and objects isn't a matter
of inserting keywords. Does it bother you that *{mi pinxe lo jubme}
would also be considered semantic nonsense, because tables aren't the
sort of thing that one can drink?

> That is, while its true that "There are no conventional
> parts-of-speech distinctions like adjectives or nouns in Lojban,"
> there are still undeniable semantic roles that we expect -- and reject
> -- from gismu when allocated to sumti placement.
>
> I think it's great that we all know from the definition that { se
> kakne } is an event/state (so it should have a nu), but I don't like
> that there are unwritten rules for which gismu have event/state built
> into them. For example, I wouldn't have anticipated the bajra/nandu
> discrepancy shown above. I wish for something like:
>
> bajra - is a process (pu'u), is a state (za'i)

No! bajra1 is the runner. You can talk about a process of something
running {pu'u barja kei}, or a state of something running {za'i bajra
kei}, or an event of something running, but none of them are something
that runs.

> nandu - is an event (nu)

nandu1 is something that is difficult to do (and anything that can be
done is an event, in lojban terms). So nandu1 is limited to events.
Similarly, nandu2 is some agent that (could or would) attempt to do
the event in nandu1. Again, putting an explicit "(event)" into the
definition of {nandu} would be good.

> See what I mean?

I'm afraid I honestly don't, and I've tried. {lo bajra} is something
that fills the x1 of "x1 runs on surface x2 with limbs x3 and gait
x4". {lo kakne} is something that fills the x1 of "x1 can perform
action x2 under conditions x3". {lo se kakne} is something that can
fill the x2 thereof. {lo kakne ku bajra} makes sense (something that
can be capable can also run), but *{lo se kakne ku bajra} doesn't (an
event someone is capable of doing isn't the sort of thing that can
run; it has no legs at all!).

[Luke Bergen]

I think the complaint that "the gismu definition doesn't somehow mark the places as needing an event/property" is kind of like saying "the gismu definition for {skari} doesn't explicitly tell me that the x3 must be a volitional entity".  It's in the semantics of the definitions that x3 of skari is a volitional entity, that the x2 of djica is an event, and that the x2 of djuno is a fact.

It's not that {djuno}'s x2 has some hidden {du'u} that gets magically inserted into it, it's just that that's the kind of thing that's expected based on the definition.  It's like saying "why doesn't it explicitly warn me that {gerku}'s x1 must be a dog for my sentences to make sense".  It's kind of silly to say such a thing.  x1 of gerku is expected to be a dog, x2 of djica is expected to be an event that is desired, and x2 of djuno is expected to be a fact (which is conveniently expressed with a {lo du'u}.  You could just as easily say "mi djuno lo fatci be fa ...", but we have a shortcut of {du'u}).

--