(A page from the Loglan web site.)
(From Lognet 96/2. Used with the permission of The Loglan Institute, Inc.)
By James Jennings
This was written before JCB's Mia memo.
I have long felt that Emerson's fio/foi model for expressing counterfactuals was more general than JCB's sio/dau/biu, but it has been rather difficult to show other people why. In this article, I'd like to show you a handful of possible systems, and show you where I think F/F and S/D/B fit into the scheme of things. A sense of perspective is often useful.
As examples, I will use variations on "If nominated I could run, if elected I will serve." (The original quote was unnecessarily complicated, soi crano.)
In a sense, all human behavior is goal driven. We eat to satisfy the goal of not being hungry. We read a book to satisfy the goal of learning something. It makes sense to have a fundamental grammatical structure that describes the relationship between two states of affairs, one of which is a goal and another which is a prerequisite or sub-goal.
"If nominated, I can run." becomes, in this model, "My being nominated is a sub-goal which is sufficient for achieving the goal of my running."
"If elected, I will serve." becomes "My being elected is a sub-goal that mandates/guarantees that I will achieve the goal of serving."
Or, in more natural English:
"My being nominated would be sufficient for my running."
"My being elected would make it necessary that I will serve."
Evaluating the probability of an outcome is central to planning a course of action. In this case, we give the probability of one state of affairs, given another state of affairs.
"If nominated, I could run." becomes "My being nominated would make it possible that I will run."
"If elected, I will serve." becomes "My being elected would make it certain that I will serve."
This model focuses on personal obligation. Given one state of affairs, what other state of affairs are we motivated through social pressure (or whatever) to strive for.
"If nominated, I could run." becomes "My being nominated would allow me to run."
"If elected, I will serve." becomes "My being elected would obligate me to serve."
Mathematical proofs often start with certain assumptions which form the context of the rest of the proof.
"If nominated, I could run." becomes "If we assume that I am nominated, then we may also assume that I will run."
"If elected, I will serve." becomes "If we assume that I am elected, then it follows that I will serve."
The "I could run" example is a little odd. When I say, "We may also assume that I will run," I mean that we don't know that I will run, but that adding my running to the set of assumptions will not contradict anything else. It's a round-about way of saying that I could run.
Before giving some examples, I need to suggest a tentative vocabulary. All of these suggested words are PAs, although some (as in the goal-driven model) might make good KOUs instead. Not all of these words will be used in the examples; ibuo ta nu pifkao lo ridle, soi crano.
Note that rui and nee are in some sense complements of each other. Perhaps one could be eliminated by compounding the other with nu (just like the kou series does) or some other variation.
This is the S/D/B model. dau and biu were borrowed from the list of discursive modifiers, and sio is a new little word.
The easiest thing to do is move the oa/oe/oi series from the UI lexeme to the PA lexeme. Thus:
There are two discursive modifiers that seem to be nice candidates. They are currently in the (UI) lexeme but could be exported to (PA). My new definitions are:
The mnemonics aren't too good for PA expansion, but perhaps they could be expanded into a complex.
All of these models have a problem in common. Although the proposed words connect two "states of affairs", they are closely bound to only one of them. The scope of the other state of affairs is determined by the PA scope rules, which can cover several sentences unless explicitly marked otherwise. We need some marker that will end the scope and reset the focus to the default. The usual scheme is to introduce yet another PA word.
Since so many of the proposed words are borrowed from the list of discursives, I'll suggest another one.
feu (fekto) : in fact/actually
I suggest using this with any of the above models.
Here are some words needed to translate "If nominated I could run, if elected I will serve."
Tiu lepo mi nu persange gu, mi nu polprusa, i, rui lepo mi nu poltia gu, mi surva.
The-event-of I am nominated, is sufficient (a sufficient sub-goal) for (the goal of) my being a candidate, (and) the-event-of I am elected, makes it necessary (is a mandating sub-goal) for me to serve.
Biu lepo mi nu persange gu, mi nu polprusa, i sio lepo mi nu poltia gu, mi surva.
Given the-event-of I am nominated, possibly, I am a candidate, (and) given the-event-of I am elected, certainly, I am serving.
Oi lepo mi nu persange gu, mi nu polprusa, i, oa lepo mi nu poltia gu, mi surva.
Given the-event-of I am nominated, I am allowed to be a candidate, (and) given the-event-of I am elected, I am obligated to serve.
Kuu lepo mi nu persange gu, mi nu polprusa, i, dou lepo mi nu poltia gu, mi surva.
Generally from hypothesis the-event-of I am nominated, we can assume I am a candidate, (and) given the hypothesis the-event-of I am elected, we can conclude that I am serving.
I don't think that these are the only possible models. In fact, you could probably write several more by adapting other UI and KOU words. (A certainty subjunctive, based on ia/ie/ii/io/iu: "Given X I am certain that Y"?) Given the wealth of alternatives, what can we do? I see four options.
If you are a sociologist at heart, you might think that the statistics of sio/dau/biu can be adapted to all cases. And it probably can through sufficiently deep metaphor. But if you were, say, a programmer/physicist, you might prefer to see everything as goal driven. And a mathematician might prefer the "given hypothesis...we can assume..." model. To choose one is to introduce a (perhaps unnecessary) metaphysical bias.
We have three kinds of abstraction, and four kinds of "because"; why not four or more kinds of counterfactual? Even the argument that it would use up too many little words doesn't seem to apply, since we are recycling a lot of the UIs for the purpose. The only reason not to adapt them all is because we've found an easier way.
Some predicates, like danza, are naturally "counterfactual" or "world creating". (Mi danza ta, I desire that, describes a state of affairs that doesn't exist, your possesion of ta.) We could make all counterfactual claims exclusively with world-creating predicates if that seemed to be the best option.
Is there a pattern? I think so. Each model relates two "states of affairs", although the actual relation is different in each case. Each model also has an "intensity scale" with two or more points on it. Therefore our "meta-model" needs to mark two states of affairs, and then specify a relation and an intensity.
We can mark the two states of affairs the same way all of the other models do: by inventing a CVV-form PA-word to mark the "conjectural" or "counterfactual" state of affairs, and letting the scope of the PA mark the "consequent" state of affairs. I'll call the word fuu for now.
Fuu lepo mi nu persange gu, mi nu polprusa, ...
Given the-event-of I am nominated (and I may not actually be), I am (in some way not specified) a candidate, ...
What predicate should we use to "PA expand" this construction? That is, what word is to fuu as pasko is to pa? That's a tricky question. Since we've designed fuu to transcend all of the models, we've also designed it to transcend all of the likely primitives. We can always (and should) make a complex for it, but I doubt anyone would find it satisfying. But I digress.
An easy way to add an intensity scale is to put a NI in front of the fuu.
I, rafuu lepo mi nu poltia gu, mi surva.
And, given-in-all-ways the-event-of I am elected, I am (in some way not specified) serving.
It is tempting to put the NI in the lepo (to get, say, rapo) but that would strongly limit the states of affairs you could describe. It would, for example, become difficult to talk about all the times you were elected (which would be rapo mi nu poltia otherwise).
The last step is to specify the relation. The only plan I have is to tack on another word or phrase. What kind of phrase is best might take some experimentation to determine. Perhaps soi, as in soi numsesmao, "speaking as a mathematician (using the hypothesis driven subjunctive)". Or maybe coi, "according to/in principle", as in coi lopo daknuu (according to statistics) would be useful. Or perhaps any of the UIs I borrowed, in their original non-PA-ed form, would be best. For example:
Rafuu lepo mi nu poltia gu, mi ai surva.
If elected, I will certainly (using an intention subjunctive) serve.
Adding another phrase is at best clumsy, so the question is, what happens if you leave off the relation specifying phrase? Can the listener usually figure it out from context? I think that I could rely on context most of the time ("If elected, I will serve" obviously seems to be some kind of intention scale, not one of the four I presented), but other logli might not be so blithe.
Why am I spending so much time on the hypothetical fuu? Well, if you assume that the relationship specifier is usually determined from context, you essentially have the fio/foi system. Fuu becomes fio and the reset operation feu becomes foi. The "relationship specifying phrase" becomes the "accessibility relation" and the "states of affairs" are then called, in the rather misleadingly romantic way of Kripke modal logic, "possible worlds".
And that is why I feel that F/F is more general than S/D/B.
Copyright © 1996 by The Loglan Institute, Inc. All rights reserved.
Send comments and corrections to:
