From: jorge@phyast.pitt.edu
Date: Thu, 24 Aug 95 19:39:15 EDT
Message-Id: <9508242339.AA15189@clueless.phyast.pitt.edu>
To: cowan@locke.ccil.org, jorge@phyast.pitt.edu
Subject: abstractions
Status: OR


Some comments on the abstraction paper. 

> On Lojban Abstraction
> $Revision: 1.10 $
...
> 4.  Property Abstractions: "ka"
...
> A pro-sumti (of selma'o KOhA) is chosen from the series "da", "de", "di".
> Typically, one that is not already in use would be chosen, but this is not
> required.  The pro-sumti is then suffixed with the the cmavo "kau" of
> selma'o UI.  (See Section 8 for more information on the uses of "kau".)

I think this convention doesn't work, because {kau} already has a different
meaning in such constructions.

> 4.9)	la djan. cu zmadu la djordj. le ka mi prami dakau
> 	John exceeds George in-the property-of (I love X).

That should mean "John exceeds George in the property of whom I love", 
which makes no sense because it leaves no place for the ones with the 
supposed property to be compared. (If I love Paul, how does that serve 
as a property to compare John and George?)

Using {ke'a} as the lambda variable,

	la djan. cu zmadu la djordj. le ka ke'a prami makau
 	John exceeds George in-the property-of (X love WHO).
	John exceeds George in loving whom they do.

	la djan. cu zmadu la djordj. le ka makau prami ke'a
 	John exceeds George in-the property-of (WHO love X).
	John exceeds George in being loved by who loves them.

Perhaps {ke'a} is not the best choice for this because it would require 
subscripts if there is also a relative clause involved (I think you once 
gave a good example of this, but I don't remember what it was now), but 
{kau} is definitely not a good choice.

> 5. Amount Abstractions:  "ni"
...
> Semantically, a sumti with "le ni" is a number; however, it cannot be
> treated grammatically as a quantifier in Lojban unless prefixed by the
> mathematical cmavo "mo'e":

I'm not sure this is true. Numbers usually can't be used in the places
where {le ni} would be used. In any case, the comparison with quantifiers
is not that relevant, if {le nu ...} is a number then it would function 
like {li PA}, a sumti, not like PA. 
 
> There are contexts where either property or amount abstractions make sense,
> and in such constructions, amount abstractions can make use of the same
> convention with "da", "de", or "di" as property abstractors.  Thus,
> 
> 5.4)	le pixra cu cenba le ka [da] blanu [kei]
> 5.5)	le pixra cu cenba le ni [da] blanu [kei]

This was not updated to the other convention.

> Example 5.4 conveys that the blueness comes and goes, whereas Example
> 5.5 conveys that its quantity changes over time.

To convey that the bluness comes and goes I would rather use something
like

	le pixra cu cenba le ka xukau [ke'a] blanu

or
	le pixra cu cenba le du'u xukau [ke'a] blanu
	The picture varies in whether or not it is blue.

I'm still not convinced that there is a quantity of being blue. I agree
that there can be variation in being blue, which is what 5.4 suggests
to me, but I wouldn't say that two different blues differ in quantity.

> 6. Truth-Value Abstraction:  "jei"
...
> 6.3)	mi ba jdice le jei la djordj. cu zekri gasnu [kei]
> 	I [future] decide the truth-value of George being-a-(crime doer).
> 	I will decide whether George is a criminal.

I don't think the x2 of jdice can be a number. What if the truth value
is .78? Does that mean that {mi ba jdice li pizebi}?

	mi ba jdice le du'u xukau la djordj. cu zekri gasnu 
 	I will decide whether George is a criminal.


> 7.3)	mi djuno le jei la frank. cu bebna [kei]
> 	I know the truth-value of Frank being a fool.

Same thing. The x2 of djuno can't be a number.


> 8. Indirect Questions: "kau"
...
> In Example 8.3, we have chosen to use "ma" as the word marked by "kau".
> In fact, any other sumti would have done as well:  "zo'e" or "da" or
> even "la djan.". 

If the semantics argument is not convincing (to me it is), at least for 
practical reasons question words should be preferred. They are more
compact and cover more ground than the indefinites or the quantifiables:

ma	da	zo'e
mo	bu'a	co'e
xu	--	--
xo	--	tu'o
ji	--	--
je'i	--	--
gi'i	--	--
cu'e	--	do'e

They are all useful for indirect questions (except perhaps {cu'e}, which
is not all that useful even for direct questions), and it makes sense 
to use always the same type of thing. 

(Perhaps some more examples with other than sumti indirect questions should
be added.)

> 12. Event-Type Abstractors And Event Contour Tenses
...
> All these cmavo are applicable to events seen as processes and abstracted
> with "pu'u".  Only processes have enough internal structure to make
> all these points and spans of time meaningful.

This section is very confusing to me, it seems to have things backwards.
I think a few examples of what is meant would be helpful. 

Only processes have enough internal structure, yes, but this has to do 
with the meaning of the selbri, {pu'u} doesn't help at all. For example,
consider the start of a process of writing. That could be {co'a ciska}.
But {le pu'u co'a ciska} is not the start of a process of writing. It is
rather a process of start writing, whatever that means. Perhaps
{le co'a pu'u ciska} would be the start of the process, but if this is
what is meant in this section, it is not at all clear.

> For events seen as states and abstracted with "za'i", the meaningful
> event contours are the spans "pu'o", "ca'o", and "ba'o"; the starting
> and ending points "co'a" and "co'u", and the achievement contour "co'i".
> States do not have natural endings distinct from their actual ending.
> (It is an open question whether states can be stopped and resumed.)

Is this talking about something like {le pu'o za'i morsi}? Or about
{le za'i pu'o morsi}? These two seem very similar, but if {pu'o} is
replaced with {co'a}, one of them makes little sense.

> For events seen as activities and abstracted with "zu'o", the meaningful
> event contours are the spans "pu'o", "ca'o", and "ba'o", and the achievement
> contour "co'i".  Because activities are inherently cyclic and repetitive,
> the beginning and ending points are not well-defined: you do not know
> whether an activity has truly begun until it begins to repeat.

So {co'a zu'o} is not well-defined?
 
> For events seen as point-events and abstracted with "mu'e", the meaningful
> event contours are the spans "pu'o" and "ba'o" but not "ca'o" (a point-event
> has no duration), and the achievement contour "co'i".

{co'i mu'e} and {mu'e co'i} both seem redundant.

> Note that the parts of events are themselves events, and may be treated
> as such.  The points in time may be seen as "mu'e" point-events; the spans
> of time may constitute processes or activities.  Therefore, Lojban allows us
> to express processes within processes, activities within states, and many
> other complicated abstract things.

Examples?


Jorge