Place structure vs. grammatical morphemes

[mashers]

coi ro do! I'm a newbie to Lojban and am just learning the basics of the grammar. The regular structure of the language really appeals to me, but I have one question. Please could somebody explain to me the advantage of place structure compared to grammatical morphemes which exist in other languages? Take the example of 'klama'. In order to understand {mi klama le zdani le panka}, I have to know that x1 of klama is the person travelling, x2 is the destination and x3 is the source. This means, as a learner, that I have to learn not only that 'llama' expresses the concept I label in English using 'come', 'go' or 'travel', I also have to learn the place structure. This is five additional pieces of information (in the case of klama). Another learner, who may also have an infirm grasp of Lojban vocabulary, may mistake the place structure of the brivla. They might think x2 is the source and x3 is the destination, meaning they misunderstand the communication.

So my question is, why is it better to have place structure files by sumti (and assume that listeners know the place structure for the brivlas you are using) than it is to just use grammatical morphemes? In this example, the English prepositions 'to' and 'from' tell the listener unambiguously that the following noun [phrase] represents the destination and source of the journey respectively.

Although this does not dampen my enthusiasm for Lojban, I would be interested to hear others' thoughts on this.

[Felipe Gonçalves Assis]

The grammatical morphemes you talk about mark thematic roles.

There is no simple unanimous way of assigning thematic roles to sumti.
If you want a small set of them, you need a lot of metaphor to fit them
to different argument places. Different languages make different choices,
which new speakers must learn.

Besides, absence of thematic roles gives us freedom to define brivla for
all kinds of relations without having to fit each sumti in a role such as
"source", "target", "actor", "patient", "direction", "frame of reference",
"instrument", etc.

The closest thing we have in Lojban are "modals", but they are very
specific, and are more used to augment place structures than to
clarify argument roles.
http://dag.github.com/cll/9/5/
http://dag.github.com/cll/9/6/

(meta: This is a recurrent theme among beginners, someone should write
a page in the wiki about it.)

mu'o
mi'e .asiz.

> --
> 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/-/IZpZYa9O-GAJ.
> 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.
>

[Lindar]

I think people forget or don't realise that the origins of the language go back to the Sapir-Worf hypothesis. =/

tl;dr (because I don't feel like writing a lot) it was intentionally made as weird as possible.

inb4 big long semantic argument/clarification; it isn't necessary to explain.

Welcome to our fair community. Stop by the IRC channel and check us out on G+/Facebook/Twitter/Reddit/YouTube.

Also be sure to check out our newest lessons at:

Lessons

Lesson Videos - YouTube

[mashers]

But those roles still exist. They are defined in the brivla. The only difference I can see is that the role being specified is not announced explicitly, it is implied by the place structure of the brivla. Surely the same effect could be achieved by having a more flexible or complete set of pronouns/prepositions, resulting in greater breadth of communication of the semantics while still allowing the speaker to be explicit about the role they are talking about without assuming that the listener is familiar with the place structure of the brivla.

[mashers]

Lindar - I agree with the fundamental principle of the language (allowing unconstrained thought and freedom of expression beyond what is possible in other languages). My question related more to the assumptions that are made when using place structure in communication, and the possible misunderstandings this could lead to.

[Felipe Gonçalves Assis]

On 13 October 2012 03:51, mashers <ma...@mashley.net> wrote:
> But those roles still exist. They are defined in the brivla. The only difference I can see is that the role being specified is not announced explicitly, it is implied by the place structure of the brivla. Surely the same effect could be achieved by having a more flexible or complete set of pronouns/prepositions, resulting in greater breadth of communication of the semantics while still allowing the speaker to be explicit about the role they are talking about without assuming that the listener is familiar with the place structure of the brivla.
>

The roles _exist_? What is the role of dunda3? Is it the same as the one of
klama2, as suggested by the English preposition "to", or is it the same as
the role of xamgu2, as suggested by the use of the dative in Latin? In other
words, is it like the destination of a movement or the beneficiary of an act?

Also, what are the roles of pritu{1,2,3}?

[Lindar]

It's just your perspective on language. Just imagine that you grew up speaking Lojban and you're looking at English. What's the purpose of "to" when it's intrinsic to the word itself anyway? I imagine a Lojban native would have trouble understanding "to" or "from". You're still thinking terms of your native language, so to you "implies 'to' already" is bad thinking. It's more like a function and less like a 'shortcut' to an English phrase.

So, in short, stop trying to relate it to English.

It's not a shortcut and it doesn't imply anything.

It's a language on its own. <3

[mashers]

I think it's natural (in fact I would argue necessary) to relate to the structures (cognitive and linguistic) which the second language learner already understands. I'm not trying to change Lojban into an English model of language. I'm simply saying that it may be less ambiguous to explicity state which role you are communicating rather than assuming that the listener knows the place structure of the brivla.

[mashers]

On Saturday, 13 October 2012 20:44:17 UTC+1, .asiz. wrote:

The roles _exist_? What is the role of dunda3? Is it the same as the one of
klama2, as suggested by the English preposition "to", or is it the same as
the role of xamgu2, as suggested by the use of the dative in Latin? In other
words, is it like the destination of a movement or the beneficiary of an act?

Also, what are the roles of pritu{1,2,3}?

 

I posted this earlier but it didn't appear. Apologies if it appears twice.

 

Yes the roles exist, but I didn't mean to imply any overlap between them. That's why I said previously that a more complete set of morphemes would be required if you wanted to speficy them explicity. So the English concept "to" would be broken down into multiple morphemes, each of which expressing one of the concepts expressed by the English "to". (I'm referring to English only as an example. Of course, the actual morphemes required would reflect the possible concepts which could be expressed within the environment, rather than within other languages.)

[gleki]

I don't understand. If English "to" has multiple meanings then why should Lojban be polysemous just like English?

But if you want one morpheme for one meaning of English "to" this is how Lojban already works.

There are also short forms of predicates. e.g. {seka'a} is kinda preposition that corresponds to the target of movement, roughly it's English "to". However, it's better to call such prepositions with a Lojbanic term "sumtcita" to avoid references to ambiguous western terms.

I must also note that {fa'a} that denotes direction of a vector and also roughly corresponds to English "to/towards" can be used in many cases.

However, in "I give this TO you" and "I go TO you" the word TO may not always denote the same relation. Therefore, {fa'a} is not a universal cure.

And lastly, I suggest learning Lojban, become fluent in it. Then we can talk about such things in Lojban itself. :)

[mashers]

On Sunday, 14 October 2012 12:22:20 UTC+1, gleki wrote:

I don't understand. If English "to" has multiple meanings then why should Lojban be polysemous just like English?

 

It absolutely shouldn't. One the most appealing features of Lojban to me is that each symbol has only one meaning. My point was not that a symbol should represent multiple meanings.

 

But if you want one morpheme for one meaning of English "to" this is how Lojban already works

 

But the morphemes aren't spoken, they are implied by the place structure. How does this work in practice if a novice has limited knowledge of the brivla place structures? And doesn't having to learn the place structure for each brivla add more learning load?

 

There are also short forms of predicates. e.g. {seka'a} is kinda preposition that corresponds to the target of movement, roughly it's English "to". However, it's better to call such prepositions with a Lojbanic term "sumtcita" to avoid references to ambiguous western terms.

 

Ah yes I remember reading about the sumti tcita. Are they used in regular conversation?

 

And lastly, I suggest learning Lojban, become fluent in it. Then we can talk about such things in Lojban itself. :)

 

I'm working on it ;) Personally I find it useful to discuss new languages metalingusitically in order to understand their internal mechanisms.

[gleki]

On Sunday, October 14, 2012 5:18:23 PM UTC+4, mashers wrote:

On Sunday, 14 October 2012 12:22:20 UTC+1, gleki wrote:

I don't understand. If English "to" has multiple meanings then why should Lojban be polysemous just like English?

 

It absolutely shouldn't. One the most appealing features of Lojban to me is that each symbol has only one meaning. My point was not that a symbol should represent multiple meanings.

 

But if you want one morpheme for one meaning of English "to" this is how Lojban already works

 

But the morphemes aren't spoken, they are implied by the place structure. How does this work in practice if a novice has limited knowledge of the brivla place structures? And doesn't having to learn the place structure for each brivla add more learning load?

How can you learn Spanish/Japanese/whatever if you speak  English?

The place structure of lojban is pretty clear. Even it's definition describe all places.

This is usually not the case in textbooks of other languages.

You have to figure out how prepositions/case-markers/whatever work in any language.

In the case of lojban there is dictionary and a textbook (only in English yet).

If I'm gonna learn another language, I will search for the same: a textbook and a dictionary.

 

There are also short forms of predicates. e.g. {seka'a} is kinda preposition that corresponds to the target of movement, roughly it's English "to". However, it's better to call such prepositions with a Lojbanic term "sumtcita" to avoid references to ambiguous western terms.

 

Ah yes I remember reading about the sumti tcita. Are they used in regular conversation?

Lojban is the most flexible language I've ever met. So I wanna use both sumtcita and brivla. They just look like different styles of Lojban.

 

 

And lastly, I suggest learning Lojban, become fluent in it. Then we can talk about such things in Lojban itself. :)

 

I'm working on it ;) Personally I find it useful to discuss new languages metalingusitically in order to understand their internal mechanisms.

Believe me (I'm not a fanatic of Lojban), if you learn Lojban then understanding the syntax of other  languages will be much easier.

[mashers]

On Sunday, 14 October 2012 14:36:12 UTC+1, gleki wrote:

How can you learn Spanish/Japanese/whatever if you speak  English?

Well, the same way I learn any language - by studying the phonology, vocabulary and grammar. Lojban is no different in that respect :)

 

The place structure of lojban is pretty clear. Even it's definition describe all places.

This is usually not the case in textbooks of other languages.

Oh I didn't mean to suggest that Lojban was in any way unclear. As I said in my first post, I love the regularity of the language and the fact that it's so well documented.

 

You have to figure out how prepositions/case-markers/whatever work in any language.

In the case of lojban there is dictionary and a textbook (only in English yet).

The literature on Lojban is definitely invaluable. I'm just getting my head around the grammar, and my main point really is that the place structure requires one to learn not just the label of the brivla, but also the positions of all of its sumti. My dyslexia is making it tricky for me to learn all of this information related to one symbol, and in my current mindset a set of morphemes which could (potentially) apply to any symbol would be easier as it would create more overlap (i.e. the morphemes could be learned once and then applied to any other symbol). I'm still playing on Memrise and am finding that I'm learning the vocabulary, but the place structure of the brivla is still not sticking in my memory (although I can remember the brivla's name (if that is the correct term)).

 

Lojban is the most flexible language I've ever met. So I wanna use both sumtcita and brivla. They just look like different styles of Lojban.

Do you think that they might emerge as different [pragmatic] styles? E.g. one a formal tone and another less formal?

 

Believe me (I'm not a fanatic of Lojban), if you learn Lojban then understanding the syntax of other  languages will be much easier.

I've read that in the Lojban literature. Could you provide some insight into how this works practically? Lojban grammar is unlike any I've come across before, so I'm interested to learn how its principles can be applied when learning natural languages. 

[Lindar]

I'd also like to point out that the set of place structures is fairly regular.

Compare: dunda, cpacu, lebna

Compare: klama, litru, cliva, plipe, cpare, muvdu

Compare: gerku, finpe, mlatu, cipni, tricu, danlu, bakni

Compare: djuno, morji, cilre

Compare: kerfa, xance, stedu, kerlo, rebla, kanla, birka, etc.

You start to see similarities. So it's not like every word has a completely arbitrary place structure. When you start to acquire more vocabulary you stop memorising individual places and start memorising what 'type' of place structure it has. I don't really remember the full place structure of {morji}, I just know that it's got the place structure for knowledge things. If it's a part of something, undoubtedly it's going to be "x1 is the $thing of x2" be it arm, leg, face, etc. If it's a living thing, it's going to be "x1 is a $thing of species/variety x2". If it has to do with goods moving then likely it's going to be "x1 = active agent, x2 = object = x3 = passive agent" and then an extra place maybe if there's something else intrinsic to the process (like {vecnu}). You can bet that if something is moving then it's going to be "x1 = thing, x2 = destination, x3 = origin" and then a bunch of other places for whatever with some sort of variation on the theme as necessary, such as bumping everything down one place to make room for the carrier in {bevri}. So instead of tagging a thematic role for every single word, the selbri have their own preconfigured and regular shapes.

[mashers]

On Monday, 15 October 2012 09:13:22 UTC+1, Lindar wrote:

I'd also like to point out that the set of place structures is fairly regular.

Good point! I'll try to look out for this when I'm learning the gismu. Do you think it's better to think about it consciously (i.e. look for the patterns) or try to learn it intuitively?

[ianek]

Watch out, because there are some inconsistencies.
See http://www.lojban.org/tiki/Gismu+list+errata and pages it links
to.

mu'o mi'e ianek

[Michael Turniansky]

On Sun, Oct 14, 2012 at 9:18 AM, mashers <ma...@mashley.net> wrote:

On Sunday, 14 October 2012 12:22:20 UTC+1, gleki wrote:

I don't understand. If English "to" has multiple meanings then why should Lojban be polysemous just like English?

 

It absolutely shouldn't. One the most appealing features of Lojban to me is that each symbol has only one meaning. My point was not that a symbol should represent multiple meanings.

 

But if you want one morpheme for one meaning of English "to" this is how Lojban already works

 

But the morphemes aren't spoken, they are implied by the place structure. How does this work in practice if a novice has limited knowledge of the brivla place structures? And doesn't having to learn the place structure for each brivla add more learning load?

 

   Yes, there is a learning load, but as Lindar points out, there is much regularity, too.  But I just want to say that perhaps you don't realize that English has exactly same type of place structure load.  Unlike, for example, Latin, where the function of a word in a sentence is communicated by word endings or Hebrew which has  a direct object indicator, so that word order isn't as importnat, English says things like "John gives Jane the ball".  The role of" John", "Jane", and "the ball" are entirely based on their position in the sentence, so are implicitly as much a part of "give' as they are of "dunda".  Lojban just makes these relationships explicit.  And as someone else pointed out, by loading up brivla with several places, that means you are also learning the equivalent of many words for the price of one.  Want to learn how to say the equivalent of the English "giver", "gift", "receiver"?  Learn just one word "dunda". 

         --gejyspa