WikiDiscuss

I have voted yes on this page. I think the definitions can be made more
precise on some points. I don't know if they need to, because there
are basically no disagreements about the meanings, but anyway here are
some more comments.

> ;noi (NOI):Incidental clause. noi is Lojban's non-restrictive
> relative clause marker. The "relative" part means that it attaches
> to a sumti to provide additional information about that sumti.

I assume a sumti is a word. {noi} attaches to a sumti to provide
additional info about that sumti's referent(s), not about the sumti
itself.

When the sumti is a quantified expression (which does not strictly have
referents) the issue is a bit more complex.

> noi
> immediately follows a simple sumti; for descriptions smuti it can appear in a
> variety of places, the semantics of which are beyond the scope of this
> definition.

It shouldn't be beyond the scope, because any complication that appears
with descritpion sumti is already present with simple sumti, which can
also be quantified.

{le broda noi}, {LE broda ku noi}, {LE noi ... ku'o broda} are
all just like {ko'a noi}.

For unquantified {LE broda}, the point of attachment is irrelevant.
The clause gives additional info about the referents of the sumti.
(Inner quantifiers don't change this.)

The difficulties arise with outer quantifiers, but these apply to
description sumti as much as to simple sumti.

When there are outer quantifiers, the CLL rule is that relative clauses
applied before the {ku} are as if they were applied to the bare
(unquantified) sumti. After the {ku} the clause applies to the
quantified sumti. So:

{PA LE broda ku noi} and {PA broda ku noi} are like {PA ko'a noi}.

{PA LE broda noi}, {PA broda noi} and {PA LE noi...ku'o broda} ignore
the outer quantifier from the point of view of the relative clause.

The "non-restrictive"
> part means that the information in the noi clause is not used to restrict
> the set of things that the sumti noi is attached to refers to. In other
> words, the noi bridi is true about the sumti noi is attached to, but
> is not necessarily enough to pick out only the things the speaker has in mind
> among all the possible things that the sumti noi is attached to could
> refer to.

I don't think the "in other words" part says the same as the first part,
and I don't think it has much to do with what noi means, but it is not
false.

> Generally, noi is only used when the referents of the sumti
> have already been explained, or are obvious, and the speaker wishes to give
> additional information.

Is that true?

> For logical scoping purposes,
> the scope of a noi clause is entirely outside the scope of the statement
> in which it is contained; its scope occurs at the point immediately after the
> scope in which it was contained ends. The noi clause should be
> considered, for scoping purposes, as occuring in its own virtual sentence
> (techinically, its own "statement" production in the formal
> grammar) after both the one in which it is contained and all further
> statements that are logically connected to the one in which it was contained.

I think that's true for attachment to unquantified sumti. When attached to
a sumti with an outer quantifier, the rules are a bit more complex. Some
quantifiers don't even provide referents for a noi clause to apply to.

> As a side effect, movement of na ku through a sentence has no effect on
> noi clauses.

When there are quantifiers involved, this is not obvious. Consider:

naku su'o broda ku noi brode cu brodi
=? ro broda ku noi brode naku brodi

If we remove the noi-clause, both sentences are equivalent.
With the noi clause, are they still equivalent? Does the
first one say that all brodas are brode, like the second one
does?

> ;poi (NOI):Restrictive clause. poi is Lojban's restrictive relative
> clause marker. The "relative" part means that it attaches to a
> sumti to provide additional information about that sumti.

Again, it doesn't really provide info about the sumti. In the case of
{poi}, I wouldn't say it provides "additional" info either, that's what
{noi} does. Anyway, it is clear what is meant, I just don't like the way
it is expressed.

> poi
> immediately follows a simple sumti; for descriptions smuti it can appear in a
> variety of places, the semantics of which are beyond the scope of this
> definition.

Again, this shouldn't be the case, because any unclarity that exists
with descriptions is already present with simple sumti. The various
points of attachment don't really introduce new complications.

The "restrictive" part
> means that the information in the poi clause is used to restrict the set
> of things that the sumti poi is attached to refers to. In other words,
> out of all the possible things the sumti that poi is attached to could
> refer to (which, for example, in the case of lo dacti is a great many
> things indeed) the sumti is actually intended by the speaker to refer only to
> those things that the sumti could refer to for which the bridi in the poi
> clause is also true.

Again, the intent is clear but I don't like the wording.

It's not about what the sumti "could" refer to but what it "does" refer
to. The sumti does have a number of referents, and the poi clause restricts
from *that* number (not from any other that the sumti in some other context
could refer to) to only those that satisfy the clause.
If {ko'a} refers to John, Paul and Mary, then {ko'a poi femti} restricts
the referents of ko'a to just those that are female, presumably just
Mary. What {ko'a} could refer to in other contexts is irrelevant.

> poi is often used with da to restrict da to
> some part of all the things which exist.

The referents of {da} are all the things that there are ("exist" is
a charged word, so I wouldn't use it here), so {da poi} does restrict to
just those that satisfy the clause.

> ;vu'o (VUhO):Long scope relative clause/phrase marker. Normally, a
> relative clause or phrase sumti binds to the last sumti to its immediate
> left, regardless of logical connectors. To have a relative clause or phrase
> bind to every member of a connected group of sumti, place vu'o after the
> sumti and before the relative clause or phrase cmavo.

Logical connectors are not special here. It applies to all connectors

> immediately after the zi'e. Using zi'e to mix poi and noi
> clauses (or pe and ne, and so on) is, for very subtle reasons, not well
> defined.

It shouldn't be too difficult to make some definitory statements about
it though. I would say that in {ko'a noi ... zi'e poi ...} the noi
clause applies to all the referents of ko'a, whereas in {ko'a poi ...
zi'e noi ...} it applies to just those referents that are left after
the poi restriction.

> || noi | PA broda noi brode cu brodi | PA broda goi ko'a cu brode .i je ko'a
> cu brodi

Works often, but not a general formula.

> poi + ro | ro broda poi brode cu brodi | ro da poi broda zo'u da ga nai brode
> gi brodi
> poi + su'o | su'o broda poi brode cu brodi | su'o da poi broda zo'u da ge
> brode gi brodi

These are correct, but don't really get rid of {poi}.

> goi, unassigned | [sumti 1] goi [sumti 2] | [sumti 1] poi du [sumti 2]

This is {po'u}, not {goi}. If sumti1 has no referent to begin with,
you can't restrict its referents to those of sumti2

> error than anything else. The winner of the no-usage prize, however, seems
> to be ge'u. This, however, seems to have been a serious error: {mi po do
> ge'u .e da} means something completely different than {mi po do .e da}, and I
> don't think anyone noticed but xorxes.

I doubt I was the only one, at least the designers must have noticed it too.
Anyway, are we really voting on whether anyone else noticed?

mu'o mi'e xorxes

__
Do you Yahoo!?
New and Improved Yahoo! Mail - 100MB free storage!
http://promotions.yahoo.com/new_mail