Re: le/lo

[And Rosta <a.rosta@UCLAN.AC.UK>]

Bob:
>    {lo} is intrinsically non-specific and {le} is intrinsically
>    specific. There's no way of using {lo} as specific or {le}
>    as non-specific. ...
>
>    > Specificity is a sometime side-effect of veridicality.
>
>    I am baffled as to how you can conclude this.
>
> Now there are two of us that are baffled!  The issue is how best to
> translate {le} and {lo} into English when using the short glosses `a'
> and `the'.

Ah: sorry. I didn't realize that was the issue. I was butting in
on your thread just to make a point about the meanings of the
Lojban gadri.

> One proposal, that I oppose, is always to use `a' for {lo}
> and `the' for {le}.

I think I'm basically in favour of using "the" for {le}, and
"a" or "some" or "at least one" for "lo".

> I am baffled because I cannot see how you fail to see that when
> translating into the English, it is sometimes (but not always) better
> to use the word `the' when referring to one object, rather than the
> word `a'.

I do see that. I hadn't realized you were trying to get me
to see it.

> Consider a `veridicality operator' that is a regular expression
> search.  In a universe with at least one pattern that it finds, it
> works like {lo}; that is to say, the search finds
> `one-or-more-of-all-the-things-which-really-match' the pattern.
>
> Let the LogFlash cmavo list of 06/01/93, 00968 be our universe: in
> this cmavo list, a search for {.ua} at the beginning of a line leads
> to two instances of the class that are veridical -- two instances that
> match the search pattern.
>
> Expressed another way, this `veridicality operator' finds the one or
> more of those that really are.  Also, note that the regular expression
> search is not specific.
>
>     (occur "^\\.ua") ==>
>
>         .uanai    UI*1  confusion
>         .ua       UI1   discovery
>
> However if you change the regular expression search to look for {.ua}
> at the beginning of a line followed by a space after the letter `a',
> then only one instance is found.  In this case, even though the
> {lo}-type operator is not specific, it finds just one specific
> occurrence since there is only one instance of the pattern in the
> universe.

with you so far, though I can't see where we're heading....

>     (occur "^\\.ua ") ==>
>
>         .ua       UI1   discovery
>
> The issue at hand is how best to translate from the Lojban to the
> English in this circumstance.

With you... BTW, I would suggest that {lo pa} is probably a good
choice in this context.

> In this latter case, is it better style to refer to
>
>     an instance found, an one pattern "^\\.ua ",
>
> or is it better to refer to
>
>     the instance found, the one pattern "^\\.ua "?
>
> If your use of English is at all like mine, you will find that the
> second translation, using "the", sounds better.  This is because the
> English phrase `an instance' suggests you are dealing with one of a
> plurality of instances, but `the instance' suggests you are dealing
> with just one instance.

I agree with you, definitely. But of course this tells us about
English, not about Lojban.

In English you have the choice between "the", which is nonveridical
but can refer to the entire extension of a singleton category,
or "a", which (I think) is veridical but can't refer to the
entire extension of a singleton category. In English, the
"singleton" criterion beats the "veridical" one, so "the" wins.

> The Lojban {lo} and {le} do not suggest
> singular or plural, which the English `a' and `the' do.

Right. (Based on actual usage, though, they do seem to generally
be used as singulars. I, though, would recommend using lo/le
for plurals, and for singulars and by default using loi/lei.)

> In this case, the Lojban {lo te facki} does not tell you whether there
> is possibly more than one instance of the regular expression:
> "^\\.ua ".  Context determines whether `the instance discovered' or
> `an instance discovered' is a better translation into English (when
> you are using short glosses rather than the longer, more accurate
> ones).

reading you loud & clear...

> Likewise, of course, {le mlatu} is quite specific, but without more
> context, you don't know whether to translate the expression as `the
> cat' or as `the cats'.  Again, it is only the context that tells you
> which short gloss to use when translating into English.

with you...

> Hence, my recommendation that English translators start out by using
> the long, somewhat unwieldy glosses rather than the short ones.
>
>         {le}    one-or-more-specific-things-which-I-describe-as
>
>         {lo}    one-or-more-of-all-the-things-which-really
>
>     lo te facki
>                 one or more of all the things which really is or are
>                 instances discovered
>
>     le mlatu
>                one or more specific things which I describe as cats

What about:

  lo gerku     "at least one dog"
               "each of some dog(s)"
  le gerku     "each of the dog(s)"

they're a bit wieldier.

--And