Semantics of modals

[Dmitry Kourmyshov]

I am not sure how modals work. I just learning the language and at first thought about posting this to lojban-beginners, but was adviced on #lojban to post it here, as the argument could be non-trivial.

The firtst two alternatives I see are:

M1. Modals add additional places to predicate, creating new predicate related to basic, but with different place-structure:

5.1)   mi viska do fi'o kanla [fe'u] le zunle

Basic predicate "x₁ sees/views/perceives visually x₂ under conditions x₃." modified into "x₁ sees/views/perceives visually x₂ under conditions x₃, the eye being x_m"

6.1)   mi viska do sepi'o le zunle kanla

The same predicate is modified into "x₁ sees/views/perceives visually x₂ under conditions x₃, using x_m as a tool"

M2. Modals introduce additional predicates, linked to the main one... somehow. Example 6.1) would be interpreted as:

"P_1: x₁ sees/views/perceives visually x₂ under conditions x₃.
P₂:using x₁ uses/employs x_m [tool, apparatus, machine, agent, acting entity, material] for purpose (du'u) P₁."

It is difficult to interpret example 5.1) in such a way, as {kanla} does not offer place to link subordinate predicate directly, as is the case with the third place of {pilno}: "x₁ is a/the eye [body-part] of x₂" — here, we only can connect x₂ of {kanla} with x₁ of {viska}, but not with whole P₁

Modal connectives seem to support this alternative:

7.1)   le spati cu banro ri'a le nu do djacu dunda fi le spati

7.5)   do djacu dunda fi le spati seri'a le nu ri banro

7.6)   le nu do djacu dunda fi le spati cu rinka le nu le spati cu banro

7.7)   le spati cu banro .iri'abo do djacu dunda fi le spati

All sentences under this interpretation have same predicate structure:

P₁ :x₁ grows/expands [an increasing development] to size/into form {zo'e} from {zo'e}.
P₂: {nu} P₁ (event/state) effects/physically causes effect {nu} P₃ (event/state) under conditions {zo'e}.
P₃: y₁ [donor] (water-type of) gives/donates gift/present {zo'e} to recipient/beneficiary y₃ [without payment/exchange].

but differ on which predicates are claimed and which are held as abstractions.

(By the way, am I right in understanding what only those modals which have short BAI form could be used in connectives?)


There could be third alternative, or at least, additional factor to consider that a recent discussion on #lojban touched. If multiple modals are present in the same statement, then the order of their appearance could matter, as they modify the main predicate one-by-one, creating scopes:

L1) {se pi'o lo forca ka'ai lo gerku mi citka}
Using a fork as a tool (accompanied by a dog (I am eating))

L2) {ka'ai lo gerku se pi'o lo forca mi citka}
Accompanied by a dog (using a fork as a tool (I am eating))

This view seem to be more easily combined with the first of my alternatives above, as ka'ai transforms {citka} "x₁ eats x₂" into "x₁ eats x₂, accompanied by x₃.", but still difference between full sencences is unclear, as they have same places. Second alternative, though, preserve the structure of scopes fully — but forces to introduce compound predicates:

"P₁ : x₁ eats/ingests/consumes (transitive verb) x₂.
P₂: x₁ is with/accompanies/is a companion of x_m1, in state/condition/enterprise P₁ (event/state).
P_1&2: P₁ & P₂
P₃:using x₁ uses/employs x_m2 [tool, apparatus, machine, agent, acting entity, material] for purpose (du'u) P_1&2."

(and vice versa for L2) )

I have to admit what view M2 is influenced by how modals are handled in gua spì and Tòaq Dzũ, but also seem to be supported, as I mentioned before, by how modal connectives work in lojban. So what do you think, which view is more correct? Or is there some other interpretation?

[Jorge Llambías]

On Thu, Jan 29, 2015 at 8:20 PM, Dmitry Kourmyshov <dmitry.k...@gmail.com> wrote:

M1. Modals add additional places to predicate, creating new predicate related to basic, but with different place-structure:

 

M2. Modals introduce additional predicates, linked to the main one... somehow.

Both are essentially correct. 

(1) A viska B se pi'o C

(2) A viska B .i jo'u A pilno C lo nu A viska B

Both (1) and (2) express basically the same relationship between three things, A, B and C, we could call it "broda": "A broda B C".

(2) is just a more expanded version than (1) in explaining what "broda" means, one that doesn't use "pi'o".  

(By the way, am I right in understanding what only those modals which have short BAI form could be used in connectives?)

All tags (I don't call them "modals" because most of them have nothing to do with modality) can be used as connectives, not just BAIs, although the meaning for some of them is unclear (e.g. ".i bau bo")

 

There could be third alternative, or at least, additional factor to consider that a recent discussion on #lojban touched. If multiple modals are present in the same statement, then the order of their appearance could matter, as they modify the main predicate one-by-one, creating scopes:

Of course, but even under M1, when scope matters you still have to take it into consideration when explaining what relationship the new extended predicate "broda" expresses.

It should be pointed out that there is no automatic one-rule-fits-all method of expanding tags, at least until we figure out what the "true" underlying predicate for each tag should be, 

mu'o mi'e xorxes

[Stela Selckiku]

.au pilno lo jbobau
(I'd like using Lojban.)

.i ku'i mi djuno lo du'u na'e certu kei du'e jundi
(But I know that too many people here aren't experts.)

.i mi te zu'e fanva
(So I'm translating.)

.i tavla fi ma'oi bai
(The subject is selma'o BAI.)

.i du'o mi jicmu tolsatci
(As I know it, it's basically inexact.)

.i ku'i lo vanbi .e lo tcaci cu vajni
(But environment and custom are important.)

.i di'e jicmu
(The basis is this:)

.i lu  broda bai ko'a  li'u
   smuni mintu
   lu  lo su'u ko'a bapli cu co'e lo su'u broda  li'u
( "broda bai ko'a"
   means the same as
   that "ko'a bapli" is somehow related to that broda )

.i tolsatci za'a da'oi ko
(Observe, this is inexact.)

.i ku'i lo vanbi cu sidju mutce
(But the context is very helpful.)

.i zo bapli .ivla me'o fe bu lo fasnu
("bapli" has an event for the x2 place.)

.i va'i ro se bapli cu fasnu
(In other words, all compelled events are events.)

.i lo tersku cu .ei racli sruma
   lo du'u lo terta'a se bapli cu du lo nu broda
(Someone being spoken to thus should rationally assume
   that the compelled event discussed is the event of broda.)

.i lo nu finti lo vanbi poi drata se sruma racli cu nandu
(It's hard to thinking of a context where a different assumption is rational.)

.i di'e troci mupli
(Here's an attempt at an example:)

.i cukta lisri fau lo cinri bai lo nundarlu se bai lo nundamba fi'o morsi du'e prenu
(A book story with an interesting event with a compelling event of arguing with
  a compelled event of fighting with too many dead people.)

.i srana .ai la .ceikspir.
(I mean for this to pertain to Shakespeare.)

.i ju'o cu'i fi'o cizra vanbi lo selfi'i be mi zo bai cu kakne srana lo na'e fadni
(Maybe, in this strange context I've invented, "bai" can be related to something extraordinary.)

.i ku'i lo fadni vanbi zo'u fadni sruma
(But an ordinary context-- ordinary assumptions.)

.i mi gleki lo nu casnu doi jbope'o
(I'm happy we're talking about it, Lojbananas.)

.i ko ko kurji
(Take care.)

mi'e .telselkik. mu'o mristeju'i

[Dan Rosén]

On 30 January 2015 at 00:20, Dmitry Kourmyshov <dmitry.k...@gmail.com> wrote:

(By the way, am I right in understanding what only those modals which have short BAI form could be used in connectives?)

The reason why CLL says that "fi'o broda [fe'u] bo" is not allowed is that this construction, together with selbri tcita, is not LR(1) which is was the target at the time.

It is expressible in PEG, and is in the experimental parsers.

[And Rosta]

On 30 Jan 2015 03:39, "Stela Selckiku" <selc...@gmail.com> wrote:
>
> .au pilno lo jbobau
> (I'd like using Lojban.)

Stela, that looks a bit glico calquey to me. I think ".au ca'a jbobau" would be better...

(I mention this because of your interest in sejbobauing in a way dictated by the inherent character of Lojban rather than by the speaker's L1.)

--And.

[Michael Turniansky]

  I think of them as in your M1 example, adding additional places to the place structure, and therefore not scoping at different levels as your L1-L2 suggests (although of course order still matters in the scope of things like quantifiers and existentials).  Roughly speaking, all modals (both BAI and FI'O FE'U forms) can fit into the English prepositional phrase template "with a ... of...." (aided by the fact that it's a very vague phrase):

se pi'o le zunle kanla (with a tool of the left eye))

se ri'a le nu do djacu dunda fi le spati (with a cause of you watering the plantf )

fi'o kanla lo zunle (with an eye of the left kind) 

   

          --gejyspa

--
You received this message because you are subscribed to the Google Groups "lojban" group.
To unsubscribe from this group and stop receiving emails from it, send an email to lojban+un...@googlegroups.com.
To post to this group, send email to loj...@googlegroups.com.
Visit this group at http://groups.google.com/group/lojban.
For more options, visit https://groups.google.com/d/optout.

[Dmitry Kourmyshov]

Wow, thank you!

But what worries me is what it is unclear which places in {fi'o}/BAI predicate correspond to the places of main predicate. It somehow parallels the vagueness of tanru.

All gismu on which connectives in http://lojban.github.io/cll/9/7/ are based have two {nu} places, and when used as connectives, it seems what connected predicates fill these places in their proper order. When {fi'o}/BAI predicate have {du'u}/{nu} place and used as tag, it would be assumed what that place will be filled with the abstraction of the main predicate; when tag predicate have an agentive place, it is most probable what that place would be filled with the agent of the main predicate — but all these are guesses and are not described in grammar.

And in examples like {mi viska do fi'o kanla le zunle}, where tag predicate have no abstraction places, it is unclear how it is connected to the main predicate — because taken literally, claim here can be "I see you. (somehow related to is, probably by common context and situation, is the fact what) I have an eye" (or even "there is an eye", as nothing indicate who or what {lo se kanla} is!).

By the way, why modals were called modals? I know what this is carried over from Loglan, and I read Loglan1, but found no explaination beyond what the word modal is used not in ordinary logical sense.


пятница, 30 января 2015 г., 4:36:43 UTC+3 пользователь xorxes написал:

[Michael Turniansky]

On Fri, Jan 30, 2015 at 11:22 AM, Dmitry Kourmyshov <dmitry.k...@gmail.com> wrote:

Wow, thank you!

But what worries me is what it is unclear which places in {fi'o}/BAI predicate correspond to the places of main predicate. It somehow parallels the vagueness of tanru.

All gismu on which connectives in http://lojban.github.io/cll/9/7/ are based have two {nu} places, and when used as connectives, it seems what connected predicates fill these places in their proper order. When {fi'o}/BAI predicate have {du'u}/{nu} place and used as tag, it would be assumed what that place will be filled with the abstraction of the main predicate; when tag predicate have an agentive place, it is most probable what that place would be filled with the agent of the main predicate — but all these are guesses and are not described in grammar.

And in examples like {mi viska do fi'o kanla le zunle}, where tag predicate have no abstraction places, it is unclear how it is connected to the main predicate — because taken literally, claim here can be "I see you. (somehow related to is, probably by common context and situation, is the fact what) I have an eye" (or even "there is an eye", as nothing indicate who or what {lo se kanla} is!).

By the way, why modals were called modals? I know what this is carried over from Loglan, and I read Loglan1, but found no explaination beyond what the word modal is used not in ordinary logical sense.

  

   Presumably because the shed light on the "mode" of the the information being conveyed by the main bridi (i.e. the how/why of it happening)

     ---gejyspa

   

пятница, 30 января 2015 г., 4:36:43 UTC+3 пользователь xorxes написал:

On Thu, Jan 29, 2015 at 8:20 PM, Dmitry Kourmyshov <dmitry.k...@gmail.com> wrote:

M1. Modals add additional places to predicate, creating new predicate related to basic, but with different place-structure:

 

M2. Modals introduce additional predicates, linked to the main one... somehow.

Both are essentially correct. 

(1) A viska B se pi'o C

(2) A viska B .i jo'u A pilno C lo nu A viska B

Both (1) and (2) express basically the same relationship between three things, A, B and C, we could call it "broda": "A broda B C".

(2) is just a more expanded version than (1) in explaining what "broda" means, one that doesn't use "pi'o".  

(By the way, am I right in understanding what only those modals which have short BAI form could be used in connectives?)

All tags (I don't call them "modals" because most of them have nothing to do with modality) can be used as connectives, not just BAIs, although the meaning for some of them is unclear (e.g. ".i bau bo")

 

There could be third alternative, or at least, additional factor to consider that a recent discussion on #lojban touched. If multiple modals are present in the same statement, then the order of their appearance could matter, as they modify the main predicate one-by-one, creating scopes:

Of course, but even under M1, when scope matters you still have to take it into consideration when explaining what relationship the new extended predicate "broda" expresses.

It should be pointed out that there is no automatic one-rule-fits-all method of expanding tags, at least until we figure out what the "true" underlying predicate for each tag should be, 

mu'o mi'e xorxes

[Gleki Arxokuna]

2015-01-30 19:22 GMT+03:00 Dmitry Kourmyshov <dmitry.k...@gmail.com>:

Wow, thank you!

But what worries me is what it is unclear which places in {fi'o}/BAI predicate correspond to the places of main predicate. It somehow parallels the vagueness of tanru.

Because a better wordlist should clearly specify to which place of the main clause it refers to.

Personally I think that it always refers either to the main clause itself ot to the x1 unless specified otherwise (and by analysing the ma'oste I found no such "otherwise" cases but that's only my formalisation).

[Dmitry Kourmyshov]

P. S.
I mixed up sematics of modal connective in M2, because places in connectives work in reverse:

7.1)   le spati cu banro ri'a le nu do djacu dunda fi le spati

7.5)   do djacu dunda fi le spati seri'a le nu ri banro

7.6)   le nu do djacu dunda fi le spati cu rinka le nu le spati cu banro

7.7)   le spati cu banro .iri'abo do djacu dunda fi le spati

All sentences under this interpretation have same predicate structure:

P₁ :x₁ grows/expands [an increasing development] to size/into form {zo'e} from {zo'e}.

P₂: {nu} P₃ (event/state) effects/physically causes effect {nu} P₁ (event/state) under conditions {zo'e}.