The singularisations

[Nick Nicholas <opoudjis@optushome.com.au>]

So where do we stand?

Where we stand is, we have getting close to a dozen singularisations 
in Lojban, each subtly different. We have had conflations between 
them, which I regard as pernicious, because they foment confusion. 
(John has been responsible for some, Jorge for others.) We do not 
want 10 singularisations in Lojban, and if we're stuck with them, we 
want ways of neutralising them by default. This may not always be 
possible.

Before I go on, reporting back from Carlson's paper on English 
Generic plurals. (I'm going through a "best of" Formal Semantics 
anthology; it makes sense for me to stick with them, since these are 
arguably the papers that have determined how formal linguists 
approach these things.) Carlson doesn't have an ontology. He doesn't 
need to, he's a linguist not a philosopher. But the way he addressed 
generics was by positing a three-way hierarchy (and I'm making this 
more coherent than he did). We have kinds of things; then, individual 
things; then, avatars of things. Kinds of things are neither space 
nor time delimited. Individuals are space delimited but not time 
delimited. Avatars are both space and time delimited. Generic claims 
are claims made of the kind, and are comparable to claims made of 
named individuals. (John walks; Dogs run.) Specific claims are made 
of avatars (John is running = The John of right now is walking; Dogs 
are barking = The Dogs of right here and right now are barking).

So for these singularisations, we'd need to be able to say whether 
they are kinds, individuals, or avatars. Or that the classes are 
inapplicable. And we'll really know where we stand when we work out 
test phrases that distinguish between the various singularisations, 
in terms of what can or cannot be claimed of one or the other.

Anyway, we have as candidate singularisations:

Set of x

Mass of x

Group of x

Mode of x (my Typical)

Stereotype of x

Prototype of x

Any-x

The any-x (I want a doctor, any doctor --- Jorge's lo'e) is really 
just a trick of scoping, though. If you're looking for a doctor, 
Lojban does already solve that for you by introducing a lambda 
abstraction: {mi sisku leka ce'u mikce}. If we could say {mi nitcu 
leka ce'u mikce} and get away with it, and the same for all gismu, 
we'd be OK. And we wouldn't be asking whether this doctor is the same 
as {loi mikce} (which it mostly is) or {lo'e mikce} (whatever lo'e 
is).

It seems that the prototype is how you recognise doctordom, if you go 
looking with the prototype doctor in mind, you'll find an instance 
(an individual of the kind.) But this doctor isn't really singulated, 
is it? You can look for two doctors quite sensibly: {mi sisku leka 
ce'uxi1 jo'u ce'uxi2 mikce} (assuming that ce'uxi1 != ce'uxi2). When 
you look for Doctor that's been through the Universal Grinder, the 
notion of two doctors is irrelevant.

(The Universal Grinder is a favoured trick of formal linguists, to 
convert individuals into masses: "there was doctor all over the 
road." If mass is distinct from group --- and I strongly believe it 
is --- then this Doctor is mass.)

That's why I still prefer something like {jaika}, {se ka}, {poi'i}, 
or whatever else is on offer. I don't buy that this doctor I'm 
looking for is singularised the way Prototypes and Masses are. And I 
also don't buy that looking for x such that x is a doctor is the same 
as walking around with a prototype in your head, and seeing if the 
things you come across are in Wittgenstein family relationship with 
it. That may be what we're doing psychologically; but if {mi sisku 
leka ce'u mikce} is legitimate, and we want a means of expressing 
that more generally, that's not what what {mi sisku leka ce'u mikce} 
is saying. {mi sisku leka ce'u mikce} is saying you're walking around 
with a lambda expression, seeing what fits it and what doesn't, 
yes-no. When you need something, you also need x such that x fits the 
slot, yes-no. You're not looking for a chimera. And prototypes are 
mental constructs and chimeras. And I believe conflating the Any-X 
(what I call this lambda thing) and the other singularisations is a 
pernicious act; even if logically it works (which I doubt), it is too 
confusing to countenance.

So: yes to being able to express the any-x succinctly. No to 
conflating it with the other singularisations on the market, and the 
prototype in particular. No to conflating it with the mass of x, 
although very frequently they will end up equivalent in context. 
Maybe to coming up with a novel gadri for it; I think it would be 
better for Lojban if we end up with something like {se ka} or {poi'i} 
instead, though, because I want to be able to see the ce'u.

Of masses and groups: when Yoko marries John, I would claim she's 
marrying the mass of Beatles, but not the group the Beatles. Whereas 
when John writes Strawberry Fields, the Beatls does write it. The 
difference, surely, is the realisation of corporate identity. The 
Mass of the Beatles is this sludge os Beatledom, which has no 
intellectual control over when to regard itself as The Beatles and 
when not. The group The Beatles is a corporate, rational entity: 
people decide whether they're writing a song as a Beatle or as a solo 
artist.

Most instances of such group/mass singularisations will not be 
rational: bits of sugar doesn't get a say in whether they solo 
artists or part of the group. Likewise, indeed, if I claim {loi remna 
cu mabru}, individual people don't het the choice of being considered 
mammals as solo artists or as part of the collective. So if this is 
the distinction between groups and masses, the default goes to masses.

I retract my Typical lo'e, since it's clearly not what even Bob and 
John wanted. I retract my understanding of squinting: squinting means 
that the singularisation is derived from the individuals, by effacing 
their differences, but clearly a prototype is a definition of the 
class, which you come into the game already equipped with. You're not 
deriving the prototype from the indivduals, you're classifying the 
individuals on the basis of the prototype.

I think the claims that {lo'e cipni} is a universal human concept, 
not a culture-specific construct, are very risky. But this is a 
popular train of thought, and people should be allowed to say it. 
Those who ideologically object to it (I'm convinced xod would, and I 
think I will) can simply refuse to use it.

I hate this. But is this-all helping?

--
**** **** **** **** **** **** **** **** **** **** **** **** **** **** ****
* Dr Nick Nicholas, French & Italian Studies nickn@unimelb.edu.au *
University of Melbourne, Australia http://www.opoudjis.net
* "Eschewing obfuscatory verbosity of locutional rendering, the *
circumscriptional appelations are excised." --- W. Mann & S. Thompson,
* _Rhetorical Structure Theory: A Theory of Text Organisation_, 1987. *
**** **** **** **** **** **** **** **** **** **** **** **** **** **** ****