Does the argument limit lead to half-ass words?

[TR NS]

In my recent studies of Lojban (and Loglan) I've started to question the efficacy of the argument system. One the things that struck me was the word for "run".

    barja x1 runs on x2 using limbs x3 with gait x4 

That seems a strange definition. I can't really think of single time I ever needed to express that the running was done with anything other then legs. Perhaps it would be useful when talking about Oscar Pistorius Olympic races, but that's a rather rare case! When I think of running, it tends to be *to* some place or at least *via* some path. The word "barja" really doesn't seem like the idea of running. It seems more akin to "treading", as in "running on a tread mill", since the definition has no arguments whatsoever for origin or destination.

But I suspect that is not what the definer really had in mind. I think rather, those arguments were left out (as if we could sensibly talk about running without them) because the definition needed to stay under five arguments and the definer already knew that a lujvo could be formed with "klama". And so we find the word "bajykla".

    bajykla k1 runs to destination k2 from origin k3 via route k4 using limbs b3 with gait b4

This word strikes me as what running is really all about. But notice we lost the surface (x2) argument. Moreover, I could easily imagine an additional speed argument.

That lead me to wonder if the ordinal argument system is really sufficient. "Running" is a concept and everything that can be reasonably associated with the concept should be accounted for in the possible arguments. While it's kind of neat how "bajykla" can be composed form "barja" and "klama", being *neat* isn't high in my list of criteria for being well defined.

On top of this, reading about Modal Tags, that really hammered home to me that the argument system has some holes. I don't see how a well defined predicate could ever make sense with dynamically added arguments. If they made sense they should already be part of the predicate's definition. (Of course, some modals are basically short-cuts for making relative clauses and not so much case tags at all. These stand out b/c they are universally applicable to just about any predicate.)

In short, it seems like the limitation of keeping the number of arguments within a small range (generally five) is an arbitrary provision that causes some concepts to be chopped-up into equally arbitrary partial concepts.  Of course, the converse issue would be how to handle predicates with potentially a dozen arguments when it is already difficult enough to recall the fourth or fifth?

[Gleki Arxokuna]

2015-05-16 5:10 GMT+03:00 TR NS <tran...@gmail.com>:

In my recent studies of Lojban (and Loglan) I've started to question the efficacy of the argument system. One the things that struck me was the word for "run".

    barja x1 runs on x2 using limbs x3 with gait x4 

{bajra} ({lo barja} is a tavern, bar)

That seems a strange definition. I can't really think of single time I ever needed to express that the running was done with anything other then legs.

I watched many videos on Youtube where people were able to run using their hands. Besides, horses don't necessarily have hands and feet.

 

Perhaps it would be useful when talking about Oscar Pistorius Olympic races, but that's a rather rare case!

So we'll use all places of this verb very seldom! Not all places are always needed.

 

When I think of running, it tends to be *to* some place or at least *via* some path. The word "barja" really doesn't seem like the idea of running. It seems more akin to "treading", as in "running on a tread mill", since the definition has no arguments whatsoever for origin or destination.

Yes, indeed. Conciseness of English definitions can sometimes lead to their incorrect interpretations.

Here is one of my examples (not sure where I got it):

xu do su'o roi senva lo ka bajra lo jdika grana lo xance be do — Have you ever dreamt of running on your hands on a narrow rod?

But I suspect that is not what the definer really had in mind. I think rather, those arguments were left out (as if we could sensibly talk about running without them) because the definition needed to stay under five arguments and the definer already knew that a lujvo could be formed with "klama". And so we find the word "bajykla".

    bajykla k1 runs to destination k2 from origin k3 via route k4 using limbs b3 with gait b4

This word strikes me as what running is really all about. But notice we lost the surface (x2) argument. Moreover, I could easily imagine an additional speed argument.

That lead me to wonder if the ordinal argument system is really sufficient. "Running" is a concept and everything that can be reasonably associated with the concept should be accounted for in the possible arguments.

I'm sure, you are not going to add places for what was the weather while the person was running or what was the political situation in Berguland at that time. All of those factors could make running somewhat different and result in different results (the weather could change the route etc.)

While it's kind of neat how "bajykla" can be composed form "barja" and "klama", being *neat* isn't high in my list of criteria for being well defined.

On top of this, reading about Modal Tags, that really hammered home to me that the argument system has some holes. I don't see how a well defined predicate could ever make sense with dynamically added arguments. If they made sense they should already be part of the predicate's definition. (Of course, some modals are basically short-cuts for making relative clauses and not so much case tags at all. These stand out b/c they are universally applicable to just about any predicate.)

In short, it seems like the limitation of keeping the number of arguments within a small range (generally five) is an arbitrary provision that causes some concepts to be chopped-up into equally arbitrary partial concepts.  Of course, the converse issue would be how to handle predicates with potentially a dozen arguments when it is already difficult enough to recall the fourth or fifth?

Just use several verbs. You can easily say {mi bajra lo jdika grana lo xance gi'e klama do ...} 

--
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/d/optout.

[TR NS]

On Saturday, May 16, 2015 at 2:22:37 AM UTC-4, la gleki wrote:

2015-05-16 5:10 GMT+03:00 TR NS <tran...@gmail.com>:

In my recent studies of Lojban (and Loglan) I've started to question the efficacy of the argument system. One the things that struck me was the word for "run".

    barja x1 runs on x2 using limbs x3 with gait x4 

{bajra} ({lo barja} is a tavern, bar)

Ha! I was wondering why the rafsi was `baj` ;-)

 

That seems a strange definition. I can't really think of single time I ever needed to express that the running was done with anything other then legs.

I watched many videos on Youtube where people were able to run using their hands. Besides, horses don't necessarily have hands and feet.

 

Perhaps it would be useful when talking about Oscar Pistorius Olympic races, but that's a rather rare case!

So we'll use all places of this verb very seldom! Not all places are always needed.

 

When I think of running, it tends to be *to* some place or at least *via* some path. The word "barja" really doesn't seem like the idea of running. It seems more akin to "treading", as in "running on a tread mill", since the definition has no arguments whatsoever for origin or destination.

Yes, indeed. Conciseness of English definitions can sometimes lead to their incorrect interpretations.

Here is one of my examples (not sure where I got it):

xu do su'o roi senva lo ka bajra lo jdika grana lo xance be do — Have you ever dreamt of running on your hands on a narrow rod?

Yea, I am not saying the limbs isn't a valid argument. It is. It's just that it seems a less useful than a "to", "from" or "via".

  

But I suspect that is not what the definer really had in mind. I think rather, those arguments were left out (as if we could sensibly talk about running without them) because the definition needed to stay under five arguments and the definer already knew that a lujvo could be formed with "klama". And so we find the word "bajykla".

    bajykla k1 runs to destination k2 from origin k3 via route k4 using limbs b3 with gait b4

This word strikes me as what running is really all about. But notice we lost the surface (x2) argument. Moreover, I could easily imagine an additional speed argument.

That lead me to wonder if the ordinal argument system is really sufficient. "Running" is a concept and everything that can be reasonably associated with the concept should be accounted for in the possible arguments.

I'm sure, you are not going to add places for what was the weather while the person was running or what was the political situation in Berguland at that time. All of those factors could make running somewhat different and result in different results (the weather could change the route etc.)

Right. The weather isn't integral to the idea of running. That's really what I am getting at. It seems like the idea of running has been broken up too much so that integral aspects of the concept have been divided across two words, not because they make sense in themselves, but just to fit a grammar limitation. 

Let me give an example in the opposite direction to clarify what I mean. Why is "klama" defined as:

    x1 comes/goes to destination x2 from origin x3 via route x4 using means/vehicle x5.

Why not instead have four simpler words for:

    x1 goes to destination x2

    x1 comes from origin x2

    x1 traverses route x2

    x1 uses/employs x2 for purpose x3

Then combine these to make a lujvo equivalent to "klama". (Note the last already exists as `pilno`.) Clearly this is just as doable as `bajykla`. And to our advantage we have words we can use without extraneous arguments -- as you said, "Not all places are always needed."

While it's kind of neat how "bajykla" can be composed form "barja" and "klama", being *neat* isn't high in my list of criteria for being well defined.

On top of this, reading about Modal Tags, that really hammered home to me that the argument system has some holes. I don't see how a well defined predicate could ever make sense with dynamically added arguments. If they made sense they should already be part of the predicate's definition. (Of course, some modals are basically short-cuts for making relative clauses and not so much case tags at all. These stand out b/c they are universally applicable to just about any predicate.)

In short, it seems like the limitation of keeping the number of arguments within a small range (generally five) is an arbitrary provision that causes some concepts to be chopped-up into equally arbitrary partial concepts.  Of course, the converse issue would be how to handle predicates with potentially a dozen arguments when it is already difficult enough to recall the fourth or fifth?

Just use several verbs. You can easily say {mi bajra lo jdika grana lo xance gi'e klama do ...} 

Sure. But I am not asking about the technical how to deal with it in the current structure of the language. Rather, I am wondering about a more philosophical question. i.e. Is "bajra" a real concept? Or is it merely a "partial-concept" that exists only because of limitations of the grammar? And if we were to make it complete, something more like `bajykla`, but with even a few more arguments, e.g.

    x1 runs to destination x2 from origin x3 via route x4 at speed x5 on surface x6 using limbs x7 with gait x8

How could such long predicates be manageable?

[Gleki Arxokuna]

2015-05-16 17:48 GMT+03:00 TR NS <tran...@gmail.com>:

Sure. But I am not asking about the technical how to deal with it in the current structure of the language. Rather, I am wondering about a more philosophical question. i.e. Is "bajra" a real concept?

Lojban by itself doesn't solve philosophical questions.

It presents options and users can invent more options. Like you may invent a new verb that will split the reality in a different way.

 

Or is it merely a "partial-concept" that exists only because of limitations of the grammar?

My opinion is that {bajra} and all other verbs are partial concepts.

And this is not just Lojban. It's just that we can't transfer the total state of the Universe in a limited amount of words.

 

And if we were to make it complete, something more like `bajykla`, but with even a few more arguments, e.g.

    x1 runs to destination x2 from origin x3 via route x4 at speed x5 on surface x6 using limbs x7 with gait x8

How could such long predicates be manageable?

They are manageable in English. At least this is what Fillmor's group thinks. FrameNet has English predicates with 7 and probably even 9 places. You learn metaphorics of english prepositions that mark places through constant usage of English.

[MorphemeAddict]

You're taking the gloss "run" too literally. The Lojban word only means what the arguments say and includes only enough of the meaning of the gloss to make it make sense. 

stevo

[Bob LeChevalier, President and Founder - LLG]

On 5/16/2015 10:48 AM, TR NS wrote:
> On Saturday, May 16, 2015 at 2:22:37 AM UTC-4, la gleki wrote:

> 2015-05-16 5:10 GMT+03:00 TR NS <tran...@gmail.com>:
>
> In my recent studies of Lojban (and Loglan) I've started to
> question the efficacy of the argument system. One the things
> that struck me was the word for "run".
>
> barja x1 runs on x2 using limbs x3 with gait x4
>
>
> {bajra} ({lo barja} is a tavern, bar)
>
>

> Ha! I was wondering why the rafsi was `baj` ;-)

bajra could have been assigned either baj or bar, but there were usually
multiple candidates for each rafsi, and the tradeoffs were complex
(probably too complex to be worth explaining at this point).

> That seems a strange definition. I can't really think of single
> time I ever needed to express that the running was done with
> anything other then legs.

bajra is intended to cover more than one kind of human locomotion, as
the gait place indicates.

> I watched many videos on Youtube where people were able to run using
> their hands. Besides, horses don't necessarily have hands and feet.
>
> Perhaps it would be useful when talking about Oscar
> Pistorius Olympic races, but that's a rather rare case!

Rare for whom? Not that rarity of expressing a given place was of prime
importance, since that is often associated with the particular language
and lexical item (and bajra is not the same as the English "run")

> So we'll use all places of this verb very seldom! Not all places are
> always needed.

Correct. All places are implicit to the concept, but context often
makes explicit usage unnecessary. Humans have only 2 legs and seldom
"run" using anything but both of their legs. We also rarely talk about
human gaits, unless they are abnormal.

> When I think of running, it tends to be *to* some place or at
> least *via* some path.

Then you are thinking of runningly-going, not running itself. To/from
are not necessarily part of running, or it would be meaningless to talk
of "running in place".

> The word "barja" really doesn't seem like
> the idea of running. It seems more akin to "treading",

"tread" is more strongly associated with walking, rather than running,
and is ambiguous between a particular kind/gait of walking and the
walking itself.

> Yes, indeed. Conciseness of English definitions can sometimes lead
> to their incorrect interpretations.
> Here is one of my examples (not sure where I got it):
>
> xu do su'o roi senva lo ka bajra lo jdika grana lo xance be do —
> Have you ever dreamt of running on your hands on a narrow rod?
>

> Yea, I am not saying the limbs isn't a valid argument. It is. It's just
> that it seems a less useful than a "to", "from" or "via".

We have "klama" and "muvdu" and "benji" for referring to interactions
between origins and destinations. "bajra" is not such an interaction,
but is an interaction between an animate "actor", the limbs being used
and the surface that the limbs are being used on (and gait, which is
more important with more than 2 limbs). bajra is thus conceptually
closer to cadzu and cpare than to klama.

> But I suspect that is not what the definer really had in mind.

I was the definer. Usefulness was not always the main consideration.
Conceptual similarities and differences from other words/concepts,
coverage of semantic space, usefulness in compounding are just a few of
the other considerations.

> I
> think rather, those arguments were left out (as if we could
> sensibly talk about running without them) because the definition
> needed to stay under five arguments and the definer already knew
> that a lujvo could be formed with "klama". And so we find the
> word "bajykla".

No. There is no magic about the number 5. For a short period, I think
it was the word pikta that had 6 places, and it made perfect sense at
the time.

There was a period where I considered adding standards places to a lot
of words, but when I tried I realized that my criteria were too
subjective, and I backed most of them out. Places were also
added/removed in contrast to other words of similar meaning.

> bajykla k1 runs to destination k2 from origin k3 via route
> k4 using limbs b3 with gait b4
>
> This word strikes me as what running is really all about. But
> notice we lost the surface (x2) argument. Moreover, I could
> easily imagine an additional speed argument.
>
> That lead me to wonder if the ordinal argument system is really
> sufficient. "Running" is a concept and everything that can be
> reasonably associated with the concept should be accounted for
> in the possible arguments.

That depends on what is essential to the concept.

> I'm sure, you are not going to add places for what was the weather
> while the person was running or what was the political situation in
> Berguland at that time.

Weather and politics are irrelevant to whether a person is running or
not; a surface and limbs are not.

> All of those factors could make running
> somewhat different and result in different results (the weather
> could change the route etc.)

I suspect you are too strongly thinking of the English word. "running"
has a meaning in politics (whether in Berguland or not), but that kind
of has nothing to do with limbs, and usually not much to do with weather.

> Right. The weather isn't integral to the idea of running. That's really
> what I am getting at. It seems like the idea of running has been broken
> up too much so that integral aspects of the concept have been divided
> across two words, not because they make sense in themselves, but just to
> fit a grammar limitation.
>
> Let me give an example in the opposite direction to clarify what I mean.
> Why is "klama" defined as:
>

> x1 comes/goes to destination x2 from origin x3 via route x4 using

> means/vehicle x5.
>
> Why not instead have four simpler words for:
>
> x1 goes to destination x2
>
> x1 comes from origin x2
>
> x1 traverses route x2
>
> x1 uses/employs x2 for purpose x3

Because it is not the case that one can have a destination without also
having an origin and a route. It is possible that you only care about
the origin or the route, because others are implicit, but they all are
part of the concept.

> While it's kind of neat how "bajykla" can be composed form
> "barja" and "klama", being *neat* isn't high in my list of
> criteria for being well defined.

My priorities were not necessarily the same as yours.

> On top of this, reading about Modal Tags, that really hammered
> home to me that the argument system has some holes. I don't see
> how a well defined predicate could ever make sense with
> dynamically added arguments. If they made sense they should
> already be part of the predicate's definition. (Of course, some
> modals are basically short-cuts for making relative clauses and
> not so much case tags at all. These stand out b/c they are
> universally applicable to just about any predicate.)

Modal tags are primarily NOT case tags, but the boundary between the two
was rather fraught (and tied up in the history of Loglan/Lojban)

> In short, it seems like the limitation of keeping the number of
> arguments within a small range (generally five) is an arbitrary
> provision

Yes, it would be.

> that causes some concepts to be chopped-up into
> equally arbitrary partial concepts.

Arbitrary in a sense, but not equally arbitrary.

> Of course, the converse
> issue would be how to handle predicates with potentially a dozen
> arguments when it is already difficult enough to recall the
> fourth or fifth?

That would certainly be a consideration. If it is difficult to recall a
place, it does suggest that it may not be necessary. But you have to be
sure you are working with the correct concept. If you are thinking
about running involving a destination, you are not really thinking of
bajra, but rather bajykla.

> Just use several verbs. You can easily say {mi bajra lo jdika grana
> lo xance gi'e klama do ...}

Lojban does not have verbs (or nouns). It has predicates. That is a
fundamental distinction.

> Sure. But I am not asking about the technical how to deal with it in the
> current structure of the language. Rather, I am wondering about a more

> philosophical question. i.e. Is "bajra" a real concept? Or is it merely

> a "partial-concept" that exists only because of limitations of the

> grammar? And if we were to make it complete, something more like

> `bajykla`, but with even a few more arguments, e.g.
>
> x1 runs to destination x2 from origin x3 via route x4 at speed x5
> on surface x6 using limbs x7 with gait x8
>
> How could such long predicates be manageable?

If they were important, they would probably be managed and therefore
manageable.

In English, verbs have an arbitrary and pretty large number of possible
prepositions that they can be linked with. In a sense, all of those
prepositions are "places" (or they might be thought of as modals or case
tags), although there is a lot of polysemy. When you learn a word, you
implicitly have to learn all the meanings of all the prepositions that
may be linked with it, yet no one ever memorizes such a list for any word.

lojbab

[TR NS]

lojbab, your reply was enlightening. Thanks for taking the time to reply.

The only thing I do not get is the gait argument on bajra. I don't see how that can be filled other then with attributive words like masno (slow) or sutra (fast). If you put a word that means "gallop" or "trot" than why would you need bajra -- you already have a word for gallop or trot.

[Pierre Abbat]

What you put in x4 is a sumti, which could be a nu-clause or a cmevla. A word
for gallop or trot, which you could substitute for "bajra", would be a brivla.
You could derive the brivla from the cmevla or vice versa, but they're not the
same.

Pierre
--
The gostak pelled at the fostin lutt for darfs for her martle plave.
The darfs had smibbed, the lutt was thale, and the pilter had nothing snave.

[Michael Turniansky]

   Just want to make sure, TR NS, that you are not under the impression that selbri can't have more than 5 places, or that is no way of referencing to the >5th places or switching them.  They can, and we can, respectively.

         --gejyspa

[TR NS]

On Wednesday, May 27, 2015 at 9:55:06 AM UTC-4, Michael Turniansky wrote:

   Just want to make sure, TR NS, that you are not under the impression that selbri can't have more than 5 places, or that is no way of referencing to the >5th places or switching them.  They can, and we can, respectively.

Thanks. My impression is the gismu are kept at 5 or less (with possible rare exception) because referencing and switching places >5 is more verbose. But I've only gleaned that from online conversation. I've read most of CLL and don't recollect reading how to do it. Where did I miss it?

[Michael Turniansky]

  Read https://lojban.github.io/cll/19/6/

  Basically, it's by using the "xi" subscript operator in connection with se or fa

       --gejyspa

[Bob LeChevalier, President and Founder - LLG]

On 5/25/2015 10:06 AM, TR NS wrote:
> lojbab, your reply was enlightening. Thanks for taking the time to reply.
>
> The only thing I do not get is the gait argument on bajra. I don't see
> how that can be filled other then with attributive words like masno
> (slow) or sutra (fast). If you put a word that means "gallop" or "trot"
> than why would you need bajra -- you already have a word for gallop or trot.

When the gismu words and place structures were created, there were no
words for gallop or trot, and I don't know them now (I don't use any of
the online dictionaries if I can avoid it) or what is in their place
structures.

We didn't design the place structures knowing what and how they could be
filled in Lojban, but rather what the word meanings required, figuring
that someone would figure out how to fill them later - I couldn't even
define the 4 legged gaits in English, much less Lojban). In most cases,
if it isn't obvious how to fill in a place, it probably requires an
abstraction sumti (LE+NU clause). If the created word for gallop has a
usable place structure, one could use "lo nu [gallop]" as a gait sumti
in bajra.

lojbab