| (being cautious) We are along way from the {su’u} thread indeed now, which
was, recall, about how to talk about the –ness or –ing of an individual in Lojban and then about what these could possibly mean. So we started with an abstract entity, assumed to exist, and asked how to refer to it in Lojban. We now seem to be talking about a well-established category of Lojban grammar, cmene, and asking some or all of the following questions. What do cmene mean? What is the sense of a cmene? How does a cmene attach to its referent? How do we pick out the right referent of a cmene (in this or any other world)? What is essential to an individual who is thereferentof a cmene? Is this connected to the cmene? Are any of these things properties or are they sui generis? What happens across worlds under the various positions on these issues? And probably a few more. Background:a world (for now and contrary to Mad Ludwig in his youth) is a bunch of things,and, as an immediate consequence, a whole bunch ofsets of things. A language is a bunch of words, and, as an immediate consequence, a bunch of sets of words, and then, related to those, a bunch of strings of words. A language is supposed to be about a world, we need some connection, an interpretation of the language in terms of a world. We start, we think, with one world and assign words of various sets to various sets of things, including words of one basic set (at least) to things taken individually. I does not matter in this process what class of words we assign to, say, simple sets, except that the grammar must somehow give rise to strings which work out to say “x is a member of s” and “s is included in t” and the like. Nor does it matter which member of this class we assign to a particular simple set, say. From another class we, equally arbitrarily,assign members to individual things, preumably from a set that allows saying that the things it points to are members and makes it at least difficult to say that they have members. [Cowan can take my talk about sets as being about discontinuous individuals, if he wants, with the corresponding kinds of relations among them.] In our initial world, a given thing will belong to some sets and not to others and will be the unique member of one set. Many of these sets will have words from the language assigned to them – or longer expressions that play the same role (as strings come to be analyzed) as words of the appropriate class. The singleton of a given object may, for example, be assigned to a word or phrase – or to several such – or to none -- in the langauge. The sets of the world form hierarchies by inclusion andsome of the higher sets may get “names” as well as the lower ones(and some at any level may not get names at all). Notice how undisciplined the connections are here: we want to say a few things but however we assign the words, we can then pick from all the strings some with appropriate structures to say this (given two names of individuals and the name of a two-place relation, any string that contains the three items will do) and, as long as we are consistent about it, it will work. Now,suppose we move to another world and suppose (it’s easier when doing this) that this world can contain things that also are in the first world, but other things as well and not necessarily all the things from the first world (indeed,not necessarily any of them – apologies to Plantinga’s ontological proof). We typically want to be somewhat less arbitrary with our language now: we know what kinds of words go with individuals, what with predicates and so on, so we will not shift these connections around (there is an alternate approach where we keep the same world by shift assignmentsor connections around, but that doesn’t improve anything but ontological muddles). Are other types of connections also more restricted – can we assign any member of the set-words class to any old set, and so on? If we look downward to the members of the set, it seems we can: sets in the new world, will, after all, likely not have the same members as any old -world set, given the the two worlds don’t have exactly the same things in them. But if we look at the level of the set and higher, we see that there are limits to this freedom. The set we call “red,” for example, in the new world must, like the set red in the old, be disjoint from a number of other disjoint sets, called “green,” ‘blue,” and so on, and fall under another set “colored” and that under “spatial,”and on up. The structure has to come over, though the particular sets are not fixed. (We would betotally lost in a world described by “Suppose red were not a color,…” though admittedly less by “Suppose a whale were not a mammal.”–the notion of “essential” in this sense is indeed scalar rather than polar.) When we come to similar questions about individuals and names, we notice we have already given the game away a bit. We talk about the same thing in both worlds before we have names for it in at least one and before we have considered what classes it has to be in (what predicates it satisfies). That is, we can identify the individual independently of what we say about it at all, and we do that because of its uniqueness, its vishesha, say, the means whereby we find the thing in any world it is in (and find out it is notin the worlds it is not in). Now we have a whole series of questionsto ask about assigning names and the like to this individual in the new world. 1. Does it have to get the same name as in the old world? Usually not: roses and the like, y’know. 2. Does it have to have the same properties – or some set of identical properties (and thus impose some further restrictions on assigning names to sets) as in the old world? Again, probably not – we can imagine everything changed in hypotheticating. 3. Does whatever gets the name thisthing had in this world in the next world have to have (some set of ) the same properties as this thing hadin this world in the next world? Still probably not – for one thing, the name may not be used at all in world 2 or not used for anything in that world at least (Cowan is a character in world 2 fiction, just as Holmes – a perfectly nice guy in world 2 – is in world 1). But further we want to be able to suppose worlds in which someone called Cowan is a master detective, without supposing the Cowan, the one we know, ever is. Somethinghas gang aglee here. It would seem that nothing could be made to follow from any hypothetical contrary-to-fact:“If Socrates were and Irish washerwoman, …” then what? The person who is called “Socrates” is world1 might well be an Irish washerwoman in world 2, but, lacking in that world all of the characteristics Socrates had in world 1, might do absolutely anything at all, without clarifying the issue the hypothetical had in mind. Similarly, an Irish washerwoman might be named Socrates in world 2 without it telling us anything useful (except about, maybe, some Irishman’s sense of humor). What we are really interested in, it turns out on careful examination is: 4. What restrictions are placed on a thing that satisfies in world 2 some description that in world 1 was satisfied bythe holder of the name? Essence,vishesha, is just numerical identity and a useful sense of a name (it solves the problem of why “Venus = Venus” is necessary while “Hesperus = Phosphorus” is not), but carries no properties with it. Onthe other hand, the name, per se, carries neither properties nor numerical identity andso is useless for most hypotheticals, which come down to laws, relations among predicates eventually. The predicate thus comes in somehow – and how else but by description? None of this makes a name ordinarily a disguised description– nor a rigid designator, for that matter. But in hyptheticating context, the (well, a) connotation of the name comes to function as its sense, the means to pick out the right person in the new world, so that we can then argue for or from some law or observation, what someone like Socrates in (often not very clearly) specified ways would do as an Irish washerwoman. So our intererest is neither in the thing nor the name, but in something two removes from either. |