[YG Conlang Archives] > [jboske group] > messages [Date Index] [Thread Index] >
This is an interesting message. Don't be put off by the length.
Xorxes:
> This is an attempt to analyse my use of the Lojban gadri
> I am not suggesting that this is how they should be used,
> I'm just trying to figure out how I do use them
Let me do the same, in this commentary. Until recently, my
usage has largely been notional -- that is, I had clear ideas
about how I would use the gadri if I did use them. I've
recently actually been writing more than the occasional
sentence or two, though.
> First, let's consider in-mind sets. In the majority
> of cases, in-mind sets are singleton sets
>
> In the case of singleton sets, le=lei=le'e. Referring
> to the members of the in-mind set one by one, or to all
> of them together, or to myopically singularize them makes
> no difference, it is always just the one member
>
> In the case of singleton in-mind sets I always use {le}
> The reason for the choice, which in principle is completely
> arbitrary between le/lei/le'e, is that {le} is the least
> marked morphologically. If I had a choice, I would make
> {le'e} the least marked and consequently use it for
> in-mind singletons, but as things stand, I prefer {le}
I use {le'e}.
I too would make {le'e} the least marked, if I could, but I
don't think we can really criticize the current assignments
for lack of foresight. With hindsight, I think that what the
designers should have done is assign no monosyllabic cmavo at
all: all cmavo should have been disyllabic. Then, once the
language had matured, monosyllabic cmavo could be assigned to
the most frequent. This would also have helped to ensure that
usage was guided more by meaning and less by mere syllable
counting. I recently been thinking about proposing to reserve
certain disyllabic cmavo as counterparts of monosyllabic cmavo,
so that a global search-and-replace can be done that makes all
cmavo of equal length (i.e. all disyllabic).
At any rate, I feel a duty to ignore cmavo length, because if
you don't it means that relatively (and unavoidable) arbitrary
or ignorant decisions made be the designers long ago are
affecting usage now, with the result that usage is not a pure
reflection of what we actually would wish to say.
> That choice of always using {le} for singleton in-mind
> categories has consequences. The first consequence is
> that when using {lei} (or {le'e}, which I hardly ever
> use) this signals a plural in-mind set. If it was a
> singleton set, I would use {le}, so when I use {lei},
> even though in theory it need not be a plural set, in
> practice it always is
For me, {le} and {lei} signal plural in-mind sets. Because there
is no lo'e/le'e counterpart of LA, I take {lai} as the singular,
because at least masses are singular and unquantified. So {la}
and {le} are the most marked.
> The second consequence is that because singleton in-mind
> sets are so frequent, when there is an occasion to use
> a non-singleton in mind set distributively, {le} tends to
> feel inadequate. How do I make sure it is not taken as a
> singleton in this particular occasion? One possibility
> would be to use {le su'o re}, but I never do that.
I would just use {le}, of course. {le za'u} is better than
{le su'o re}, if only because it makes plurality less arbitrary
({su'o re} makes it seem no more special than {su'o ci}). I
would use {za'u broda} for plural nonspecifics.
> In most
> such cases I use {ro le}. In theory {le} = {ro le}. In
> practice I never use {ro le} for {ro le pa}, it would
> sound strange, so {ro le} normally implicates {le su'o re}
> Sometimes I do use plain {le} for non-singleton in-mind
> sets, but this is rare. I also find it rare in the usage
> of others. If {le'e} were my default choice for singletons,
> then the use of {le} would automatically signal plural in
> the same way that {lei} does
Good for you!
> I'm not sure why I almost never use {le'e}. It may be that
> I use {lo'e} sometimes when I should use {le'e}, or that
> it is just not a very frequent meaning. I don't know
In my Earnest translation so far:
total lo'e (loi'e) 208
before du'u/ka 102
elsewhere 106
lo 48
loi 6
le'e (lei'e) 61
le 2 (both {su'o le}
lei 9
lo'i 3
le'i 0
Pretty interesting, eh?
> Now what about {lo'i} cases? Unlike the cases for le'i, most
> lo'i are not singletons.
True in terms of sets of things in the universe, but surprisingly
not true in terms of sets of thing we talk about. (E.g. {lo'e do
mamta}.)
> In the exceptional cases when they
> are singletons, we have that lo=loi=lo'e(=le=lei=le'e). They
> are equal to the {le} cases if we ignore veridicality. There
> is no specificity distinction possible when lo'i is a singleton
>
> So, again, when {lo'i} is a singleton I keep using {le}
> In principle, any of the six articles would serve for
> these singleton categories. But, given that singleton
> categories are very frequent in in-mind sets and rare in
> general sets, it makes sense to choose one of the in-mind
> articles for this function and use it throughout, unless
> for some reason we want to emphasize that it is intrinsically
> a singleton category in which case {lo pa} (or {loi pa}, or
> {lo'e pa}, it makes no difference) would be appropriate
I use lo'e here. I don't use lo or loi, because they imply that
the distributive/collective distinction is pertinent. I don't
use le'e/le/lei, because they imply that I'm selecting some
particular subset of lo'i.
> lo/loi/lo'e implicate a non-singleton lo'i in the same
> manner that {lei} does, just because I would be using
> {le} if lo'i were a singleton. So I never say {lo stedu
> be mi} if {lo'i stedu be mi} is a singleton. Even though
> in principle I could say it, I always prefer {le stedu
> be mi}. The consecuence is that {lo broda} and {loi broda}
> always implicate an underlying {lo'i su'o re broda}, and
> also {lo'e broda} implicates either {lo'i su'o re broda}
> or eventually {lo'i no broda}, if we accept that there are
> intensions with no extension
This makes sense if {le} is your default. As it is, I agree on
everything except {lo'e} -- since {le} is not my default, it
makes sense to use {lo'e}, the gadri for singletones, when I
am referring to a singleton.
> This also justifies the use of {le du'u}, (at least if
> we interpret that {lo'i du'u <bridi>} is always a
> singleton)
Indeed so. And likewise I use {lo'e du'u}.
In summary, my usage is better if you studiously avoid considerations
of morphological markedness. Your usage is better if morphological
markedness is an issue. We totally agree on everything but this
issue.
I think it is better to ignore morphological markedness when we are
road-testing the design, exploring it and finding ways to optimize
it within the constraints of the baseline.
In the long run we of course want a language where morphological
markedness exactly parallels semantic/discoursal markedness.
--And.