From Lognet, issue 95/1.
Two logli have just earned Master’s degrees using Loglan. Steve Rice used L for his thesis in technical writing. The result was his book Understanding Loglan, the “Loglan Zero” primer we have needed to hand newcomers for so many years. When The Institute publishes it—and we are preparing it for publication now—it will be titled Loglan 3: Understanding Loglan, and will appear first serially in La Logli and later as a book. The second logli who has used L in his academic work is Alan Gaynor. In his Master’s Thesis, Alan showed that L is capable of serving as the language of an expert system he’d designed that he calls “The Philosopher’s Assistant”. As if that weren’t enough, our Lodtua, Randall Holmes, who teaches in the mathematics department at Boise State in Idaho, mentioned some months ago that he had been funded by one of the national funding agencies—I don’t recall which one—to develop a “theorem-prover” that will do things that other computer-based theorem-provers have not yet done. (We must get him to explain that to us one day.) The L-relevent news about this project is that Randall is exploring ways of using L as the language in which to write his theorem-prover. In short, L is proving useful for all sorts of intellectual exercises...even if it isn’t making The Institute rich by popping up in lunchboxes!
* * *
It’s one of the fundamental doctrines of Loglan that each new usage we adopt be provided with an “interpretation.” Usually that’s nothing more than a rule for expanding any instance of the given usage into some logically explicit, and usually much longer, expression. My favorite example is the “observative” use of lo. Here the rule is that any free-standing utterance of the form Lo <preda>!—e.g., Lo fagro! (Fire!) or Lo simba! (Lions!)— may be translated into the much longer expression Mi na sanse lo sanpa je lo <preda> = I am now sensing signs of predas, e.g., lions, fires, snow on the ground, jobs well done, etc. The requirement that we always be able to provide such an interpretation is just as important for maintaining our logical language in truly “logical” condition as preserving the syntactic unambiguity of our grammar. And there is an even older tradition that we also try to observe—one equally important for L’s logicality, I suspect—namely that whenever we introduce a new operator X we provide a procedure for “eliminating” X by showing its equivalence to some older, X-free form. Presumably the older operators are also eliminable in this way.
As an example of the eliminability principle at work, Randall Holmes recently proposed that we adopt three new “reflexive” predicate operators—one of the words adopted was nuo; see LN 94/3:18—that would allow the hearer to infer that whatever argument occupied the first place of the nuo-ed predicate would normally occupy its second place as well. Thus, John loves John, which is normally rendered in E by John loves himself, had always been rendered in L by La Djan, cluva la Djan or La Djan, cluva Dai; but with nuo, the alternative form La Djan, nuo cluva (John self-loves) would also become available. Nuo makes a particularly good example of an “eliminable operator”; for its elimination is extremely easy. Any sentence containing nuo, say X nuo preda Y Z ... W, can always be rewritten in an equivalent nuo-less form, in this case, X preda X Y Z ... W. So anyone who wishes to write or speak L without nuo can easily do so.
If every L operator can be eliminated in this or some other way, as I believe each can, then there exists a version of L in which nothing but predas, non-designating variables, and the logical connectives will appear. No human would ever want to use such an operator-free language; it would be exceedingly long-winded. But if such a version of L exists even in principle, then this property would demonstrate L’s potential use a logic theorem-prover...just as, in fact, our Lodtua is planning to use it.
* * *
Somewhat richer examples of eliminability are provided by the simple tense operators pa na fa. Let’s consider them first in their “prepositional” roles: X preda pa Y = X predas before Y. What does this mean? Clearly the claim it makes involves both the predicate pasko (is before) and the event-operator lepo: Lepo X preda ga pasko Y = The-event-of X being-a-preda is-before/occurs-earlier-than Y. Similarly, nadzo and futci provide “explications”, as they might be called, of the prepositional senses of na and fa. But what about the adverbial senses of these tense operators? Well, don’t we assume that ti—taken to designate “this”/”this place/moment of speech”—is the implicit operand of pa used alone? Isn’t X preda pa? short for X preda pa ti? If it is, then the so-called “adverbial” uses of PA are implicit prepositional uses with the operand ti implied. Finally, the “inflectional” uses of PA may also be eliminated in favor of these predas. For isn’t X pa preda equivalent to X preda pa? This is a brief account of how the simple tense operators may be entirely eliminated from the lan-guage. Of course, the procedure outlined leaves us with the compound operator lepo to explicate. But lepo is also eliminable. Perhaps you’d enjoy working out an elimination procedure yourself.
* * *
James Jennings discusses case tags in a major article in this issue (“Toward a Theory of Case Tags”; p.16). He proposes that our understanding of the semantics of L’s predicates can be deepened by studying their case patterns. To support this, James explains Lakoff’s thesis that cases and their tags are fundamental features of human languages. It’s an interesting thesis; but it seems unlikely to be true. Despite Lakoff’s indubitable contribution to our emerging understanding of human “mentalese”—a phenomenon attested to not only by cross-cultural but also by cross-“taxonal” (if you will permit me that neologism) constancies in our perceptions of the world (see Alan Gaynor’s article in this issue)—his hypothesis about the universality of linguistic cases doesn’t impress me. Too many languages have no sign of cases. What they have in common that Lakoff evidently has documented are some strangely similar styles of metaphor, which may indeed reflect a common mentalese. As to the utility of case analysis in L, I believe our experience with cases—or perhaps I should say my experience with them, for we may each have had different experiences—can lead us to the opposite conclusion. Case tags, as well as the cases they represent, seem to be quite dispensable features of L...nearly useless ones, in fact. Although well described and publicized when experimentally introduced in 1987 (see NB3), they remain almost unused. Have a look at the texts that have been published since 1987...at the Lo Nurvia Logla columns in LN, for example. Case-tags are about as common as hen’s teeth in them. I was and am surprised at this; but already the experiment seems reasonably conclusive. Perhaps improving case-tag assignments in the ways James Jennings is now advocating would increase their use; but I remain skeptical. Case tags were introduced to solve certain problems: mainly incompletion and keeping up with the suteri places of long predas. But those problems seem not to exist any longer! Or are being differently solved.
My own experience with case tags suggests some of the neat ways we do this. My perhaps increasingly loglaform mind, soi crano, seems actually to rebel at using case tags...and I never seem to need them. If I want to say X talked about Y, for example, I immediately reshape my E intention by inserting “to someone.” So I write Da takna ba de (X talks to someone about Y) immediately. It never even occurs to me any longer to “solve the problem of the missing case” with the case tag neu. (Da takna neu de!!!! Ugh, my mind says, how ugly! Perhaps my brain really is becoming more loglaform!) In short, incompletion is a non-problem for me. So are long predas. If I want to say that X is a sign in situation Y—skipping over three places of this long preda to even think that!—I simply write Da sanpa ba be bo de, which actually says (and much more elegantly than I can in E) that X is a sign of something to some interpreter that disposes that interpreter to some action in situation Y. How satisfying that I can say all that so briefly! And without leaving out any essential ingredient of the sign-relationship! The one practical task for which I would use case tags (if I did much of it), is translating texts from preposition-rich languages. To capture intent in such languages, case tags may in fact be indispensable.
In short, my own conclusion from these years of L-watching is that the solution to nearly all the problems that case tags were originally intended to solve is to learn the place-structures of your predas. It’s not easy, but it’s possible. And it works. (We probably need to write a new teaching program to help us do that more efficiently. Any takers?) And of course this doesn’t mean that we shouldn’t improve the memorability of our place-structures whenever we can... and that includes using James’s “case theory” insights.
Here’s another approach. The order of arguments in sutera predas is often an unpleasant mystery, one that drives us these days to consult LOD “just to be sure”. Suppose we were to attempt to find—or to write and impose—an “ordering” algorithm that we could set to work on the places of predas...especially the sutera ones. If we could find such an algorithm, and if it were simple enough for humans to learn, then the mystery would go out of sanpa, vedma, and ketpi. Any takers for this one? The algorithm itself, of course, would make ample use of cases...and so of James Jennings’ insights into these matters.—JCB