(Originally appeared in Lognet 97/2)
In Lognet 96/2 I wrote an article called “Counterfactuals in Perspective”. In it, I tried to show how all of the models that I could think of for a Loglan quasicounterfactual/subjunctive marker could be seen as special cases of the Fio/Foi system proposed by Emerson Mitchell and described in my article “I Would If I Could” published in Lognet 96/1.
In that same issue, James Cooke Brown had an article called “The Mia System of Subjunctives”. I had not seen it at the time I wrote my article so it had not included the Mia system. The question remains, is mia a special case of fio? My gut reaction was yes, it is, but I could not see how they were related until recently.
Let me start by summarizing the two systems.
The Fio system is described with a geography metaphor. The states of affairs under consideration are metaphorically points on a map. The ability to go from one state to another is a road connecting two points on a map. The central idea is “accessibility”: given a starting point and a set of roads, what other points can you get to? This is a very general mathematical thing to do. Whenever a mathematician talks about a function mapping the elements of one set onto another, they are using this metaphor.
Here is an example using the metaphor literally.
If I were at the bank, I could go to the market.
I could translate it literally:
Kanoi mi nu sitfa le banko, ki mi kanmo lepo godzi le marte.
If I am located at the bank, then I am able to do the event of going to the market.
But let me make it a subjunctive claim. Here the “states of affairs under consideration” are points on a map. I am claiming that “going to the market” is accessible from the bank...that there is a path connecting them. It is this path that is important, and in the Fio system it is indicated with the word fio.
Fio lepo mi nu sitfa le banko guo, mi godzi le marte.
In a “possible world” where I am located at the bank, my going to the market is “accessible”.
There are two differences between the literal and subjunctive versions.
The literal version uses a logical if. That means that it is only false if I am at the bank and I am unable to go to the market. If I were across the street from the bank and bound hand and foot, I would not be able to go to the market, even if someone carried me to the bank, and yet the claim, If I were at the bank I could go to the market would still be true, which is not usually what people expect.
The subjunctive version is more vague. It doesn’t say, as the literal version does with kanmo, in what way the “at the bank” state of affairs is related to the “go to the market” state, merely that the latter is accessible from the former. It assumes that [the phrase] is able to is underst[andable] from context.
The Mia system is based on an imagination metaphor. The state of affairs under consideration is considered to exist in the speaker’s mind isolated from the real world. [Not quite. A more exact way of saying this is that speakers’ minds are isolated from the intersubjectively observable world. The unique thing about people’s minds is that each mind is observable by exactly one privileged observer, the person whose mind it is. That may not always be true, but it is true now.—JCB] In an example from JCB’s article:
Mi hapci mia lepo mi bragai.
I am happy inanimaginaryworldinwhich I am king.
I would/could be happy if I were king.
Just as in the fio example, we could write this using the logical if:
Mi hapci, icanoi mi bragai.
and it would have the same logical problem. It would be true if I am not king, even if kingship would in fact make me miserable. Unlike the Fio system, Mia tells us in what way the two states of affairs are related—they are “in the speakers mind”. Actually, it does not describe two states of affairs as fio did. It describes the entire situation being imagined as one state of affairs. [Not true. It describes a cause producing an effect, both of which are in the speaker’s imagined world.]
There is one other difference between the metaphors these systems use. As Emerson Mitchel wrote, “Fio is usually explained using ‘alternative worlds’ metaphysics but this is not an intended feature of the defining predicate. Rather it is an accident of the best English explication.” In other words, although we use the geography metaphor to explain the Fio system, we are to resist defining it in terms of that metaphor. It is this, in part, that gives the Fio system the power to embrace all of the other systems. The Mia system, on the other hand, actually adopts the metaphysics of its metaphor. [What it adopts is the physics of its metaphor. Brains, after all, and their neural contents and happenings, physically exist.]
The Mia system states that if we don’t wish to specify the kind of world being imagined we use an indefinite variable.
Mi hapci mia ba.
I amhappy in some unmentioned imaginary world x.
I could/[should/]would be happy.
Although it hasn’t been mentioned in any article so far, I see no reason why the Fio system couldn’t do the same thing.
Mi hapci fio ba.
“I amhappy” is accessible from some unmentioned “possible” world(s) y.
I could/would be happy (given unmentioned conditions.)
The Fio system states that if you leave off the argument to fio entirely, as when using it as a “tense” marker in the next example, then the assumed world is the real one in which the speech is taking place.
Mi fio hapci.
“I am happy” is accessible from this, the real world.
I could/would be happy (given the current conditions.)
JCB’s [original] article doesn’t mention what the Mia system does in this case [but I do mention it in my second article, the one in LN 97/1: 5ff; in particular, see p.17, col.1, Sec.1], but it can’t make the same assumption without defying its own metaphor.
? Mi mia hapci.
I am happy in the imaginary real world. (???)
Perhaps it only means that I could be happy if I only imagine that I am. Or that I only imagine that I’m happy. I can’t really make sense of it. [Try putting it in a speech made by some Alice in her Wonderland.]
Both systems have a way of adjusting the strength of their claims. This allows both systems to map their claims onto ... English would[, should,] and could, and both systems have generalized it a bit.
The Fio system uses the fact than any PA word (which fio is) can be compounded with any NI word. This gives a very fine grained set of distinctions, and is described in detail in the article “I Would If I Could” (Lognet 96/1: 5ff). I give only three examples here.
Mi hapci rafio lepo mi bragai.
I am happy (is accessible) in all possible worlds where I am king.
I would be happy if I were king.
Mi hapci rofio lepo mi bragai.
I am happy (is accessible) in many possible worlds where I am king.
I probably could be happy if I were king.
Mi hapci sufio lepo mi bragai.
I am happy (is accessible) in at least one possible world where I am king.
I (barely) could be happy if I were king.
The Mia system adjusts it’s strength by compounding mia with the ... “likelihood” operators, sio, dau, and biu for certainly, probably, and possibly. The result is similar to the three examples above.
Mi hapci siomia lepo mi bragai.
I would (certainly) be happy if I were king.
Mi hapci daumia lepo mi bragai.
I should (probably) be happy if I were king.
Mi hapci biumia lepo mi bragai.
I might (possibly) be happy if I were king.
Note that the Mia system treats the “likelihood” distinction of would and could separately from the “imagination” part, while the Fio system uses a natural part of its metaphor, the number of geographical points, to make the same distinctions.
In particular, mia is not the equivalent of either would or could. It is the “imagination” marker, and the “likelihood” markers are added separately. In contrast fio, as proposed, is roughly equivalent to sufio and can often be translated as could.
I would like to propose an amendment to the Fio system that would amend this disparity. I suggest that fio be numberless in the same way that le is. Le mrenu means the one or more men I have in mind, so fio would refer to the one or more worlds I have in mind. In this way, a bare fio wouldn’t make a distinction between would and could; you’d need sufio and rafio for that.
Just as pa and fa have a cousin, na, that returns you to the present time, the Fio system suggests a word, foi, that returns you to the real world. It is important to remember that this is its main function, and that in this function it appears without an argument.
Mi fio hapci, ibuo mi foi no hapci.
I could be happy, but I’m not happy.
The Mia system does not have an equivalent “reset” operator. The Mia proponents don’t seem to think it’s necessary, while the Fio proponents seem to think it’s vital. (Curiously, it is this small feature that is the link between the two systems. See below.)
Although foi’s purpose is fulfilled without ever giving it an argument, the fact is that since it is a PA word, you can give it an argument. What does it mean then? Well if
Mi hapci foi. means
I am happy in the real world. then
Mi hapci foi lepo mi bragai. means
I am happy in a world where I am king.
At first blush this looks to be the same thing as
Mi hapci fio lepo mi bragai.
but it isn’t. The accessibility is gone. The central idea, the map that connects states of affairs, is missing.
I have often translated foi using would because, when it is lacking an argument, it refers to all the real worlds (there’s only one) and has the wouldlike impact of rafio. It’s as if Mi hapci foi. means, I would be happy in the real world (and this is the real world, therefore I am happy.) When foi does have an argument, translating it as would probably isn’t always justified since foi, like fio, should not specify a particular number of possible worlds. (On the other hand, rafio without an argument, which means in all worlds accessible from this world or definitely, in this world could be used as a “reset” operator. Perhaps we should consider ditching foi in favor of rafio.)
Foi with an argument is amazingly powerful considering that it was just an afterthought. You could craft a wide range of subjunctive statements with just foi, even getting variants of would and could by compounding it with sio, dau, and biu, but it isn’t as flexible as fio. An argumentless foi is not subjunctive at all, while an argumentless fio is, and can refer to an undescribed “possible world”.
And remember that it was the accessibility relation that allowed fio to stand in for all of the different subjunctive proposals in the “Counterfactuals in Perspective” article.
Fio: Geography (used to explain) —Relates two states of affairs.
Mia: Imagination (part of the metaphysics) —Relation is encapsulated in one, imagined, state of affairs.
Fio: Refers to the “world” of the speaker.
Fio: Compounded with NI. (Defaults to su, but the proposal can be changed to be more general.)
Mia: Compounded with sio/dau/biu.
Fio: Defines a new word, foi.
Mia: Issue is not addressed.
What is the Mia system, really? I originally tried to interpret the main word of the Mia system, mia, in terms of the main word of the Fio system, fio, but it didn’t work. Fio’s ability to relate two states of affairs was contrary to mia’s insistance on putting everything in the mind of the speaker. But foi, which puts us firmly in one world, is a lot more like mia, which puts us firmly in one imagined world.
Let’s look at the PA eliminators.
Mi hapci mia lepo mi bragai. becomes
Ba minsia lepo mi bragai, guo lepo mi hapci[, guo mi].
Somethingx isanimaginaryworld inwhich my beingking (causes) me tobe happy[, and I imagined it].
Mi hapci foi lepo mi bragai. becomes
Lepo mi hapci, guo feorkui lepo mi bragai[, guo] ba.
My happiness is factuallyrelatedto (occurs in the same world as) my kingship according to some pointofview/observer/metaphysics/reality x.
To bring these two into line, all we need to do is specify the pointofview [argument] of feorkui. If we want to have the same meaning as mia, we need to replace some x (ba), with in my mind/imagination (lomi smina sitfa, or lomi po tcupeo.)
I am not claiming that foi is mia. They have very different backing metaphors. I am claiming that they have similar usepatterns, and that foi can stand in for mia when you specify the metaphysics. It is interesting to note that the Fio system can do completely without foi with an argument, while the Mia system doesn’t have a use for mia without an argument. It’s as if the systems are grammatical complements of each other.
JJ has made a useful contribution with this paper to our understanding of the two subjunctive systems that have been under development in the Loglan community for the past three years. Since p was written, however, the Mia system has been formally adopted by the K; what is here called the Fio system was formally rejected as s’s only rival; and, along with mia, one of the essential elements of the original Fio/Foi system, namely foi, was adopted as one of the five auxiliaries that turned out to be necessary to disambiguate the English ingredients of the Mia system. So foi is now the verbal auxiliary that conveys the objective sense of both L oa and E must, having been derived from folfunrui = is obliged to; see Lognet 97/1: 10-13 for the details.
These adoptions were announced in the Sau La Keugru that appeared in that Lognet in March of this year. Some time later Emerson Mitchell and James Jennings submitted a new subjunctive proposal to the Keugru, presumably to be adopted alongside Mia and used by those who wish to use subjunctive forms to make “objective” claims. The words kui/feu were to be the critical operators of the new system; and, while the new system is similar to the Fio/Foi one in many respects, it does have some differences that recommend it to the K as a possible alternative to Mia. We have decided to try Kui out provisionally as a possible comrade of Mia, although the protocols of the actual experiment have still to be worked out.
To be quite clear, therefore, just what the Mia system provides, I feel obliged to correct the several misapprehensions about it that JJ has managed—I’m sure quite accidentally—to convey in this earlier article of his; but, to be fair, it is one he wrote before the adoption of Mia was announced last March. (Sometimes papers lie around on disk until we find room for them on paper.) I took the liberty of trying to correct some of these misapprehensions in situ; others I have saved for correction in this commentary.
Certainly the most far-reaching of these misapprehensions is JJ’s presumption that the designers of the Mia system believe that “minds”—which are indeed the locations of the subjective worlds with which Mia deals—are not part of the “real” world. But given the scientific model under which L, including Mia, has been developed, this is not true. In that model—one largely derived from the work of the American philosopher/psychologist George Herbert Mead; see his Mind, Self, and Society, 1934—brains, and the mental activities taking place in them, are as much a part of the physical world as sticks and stones. What distinguishes brains and their resident minds from many other parts of the physical world, according to that model, is that they are not accessible to the sort of intersubjective observation on which science (currently) depends. Each mind has only one privileged observer: the person whose mind it is. This puts Mia statements, which are about happenings in these secluded regions of the physical world, in a peculiar epistemological position. They cannot (yet) be either verified or refuted by the methods of science.
This does not make them useless...quite the contrary. It is very useful to know that one’s interlocutor is telling you about something that only i can see. Coupled with skepticism, such knowledge warns one not to take i’s claims too seriously...or, with a warm faith in i’s honesty, i’s subjunctive remarks can provide a sympathetic listener with important information about i i-self. (Remember the old lady on the bus? If you believe her, how nice to know that she is of that magnanimous kind!) But what treating one species of subjunctive claims subjectively does for us linguistically is to rid at least that species of the metaphysical baggage with which so many other senses of the subjunctive are heavily burdened. Surely that is a useful move in a logical language!
But this move also prepares the way for a companion operator, for an allolex of Mia—the one I’ve called “xxx”—in which some metaphysical presumption actually is on the table. No L-speaking spiritualist, or, for that matter, cosmologist, would be content without one. That’s why the current Mitchell- Jennings proposal to provide us with an allolex—or rather, to be honest, with a sort of an allolex—interests me so greatly. Keep tuned. We are not through with the Loglan subjunctive yet!
The second most critical piece of misinformation that I found in JJ’s article—I’m sure it was unintentional—is that Mia fails to relate the subjunctive outcome to the subjunctive condition due to its “encapsulat[ing these two events] in[to] one, imagined state of affairs”. JJ contrasts this with “Fio’s ability to relate two states of affairs”, which, he says, is “contrary to Mia’s insistence on putting everything in the mind of the speaker”. But these remarks are surely misleading! Mia plainly differentiates the imagined cause from the imagined effect by always putting them in either the E mia C order or the Mia C, E order. Thus E mia C claims that effect E takes place in a world imagined by the speaker in which cause C also takes place: I am happy in an imaginary world in which I am king. With Fio, we do substantially the same thing. We say W1 fio W2, although here the relation between the the two “worlds”—indeed, they are like places on a map—is merely their connectedness, the “accessibility” of one place from another.
From JJ’s exposition of these matters, one gets the impression that accessibility is a weaker relation than causation. This is a difference, alright; but it is a difference between one relation and another, not between the presence and absence of differentiation of the related events.
The point about the Mia system’s causal claim is that, even though subjective, it is definitely causal. Causation requires that the presence of the (set of) causes be followed by the presence of the (set of) effects; that their absence be accompanied by the absence of the effects; and that any variation in the strength of one set be accompanied by variation in the same direction in the strength of the other. (All this was settled years ago by the British logician John Stuart Mill, and has been part of the scientific canon ever since.) Accessibility, as I read JJ, apparently only requires joint presence. But the causal laws used by the speaker to relate s’s imagined effect with s’s imagined cause, under the Mia interpretation, are, in most cases, simply extensions of the causal laws s imagines to be at work in the intersubjective world outside s’s head. The Fio system seems not to welcome inquiry into these details. There seems to be no place for them in its eliminative predicate (though this could be repaired). In claims made with the predicate minsia, however—with which we eliminate mia from its claim—there are places for all these objects of inquiry.
A third and fairly trivial item of misinformation is JJ’s suggestion that Mia somehow cannot handle “missing conditions”. Not true. The accommodation of the missing-condition subjunctive—e.g., I would be happy (sometime, somewhere)—was first made in the Sio/Dau/Biu system, a direct ancestor of the Mia system. For Mia, mia ba now plays that role quite satisfactorily: Mia ba mi hapci. Under some conditions I would be happy. I am glad to see that JJ sees this move as worth adopting for the Fio system.
This brings me to the somewhat extraneous role of Foi in the old Fio/Foi system, a role evidently now to be played by Feu. Clearly, foi is a “yank-back” operator: it yanks us out of the subjunctive mood and puts us back in the indicative again. JJ is quite right to point out that the architects of Mia have so far not found a need for a specifically subjunctive version of such an operator.
We of the Keugru are aware, however—and have been for some time—that adding some yank-back feature to our present “tense” system (and Mia is part of that system) will probably be necessary. But we haven’t discovered yet whether there should be several specific versions of these yank-backs—say, one for time, one for space, one for causation, one for mood, etc.—or whether just one yank-back will do for all.
This sounds like a simple question to answer but is in fact an fairly complicated one. But I promise that we of the Kuegru—aided by any other tiftua we can enlist to help us (please let us know if you’d like to do so)—will study the yank-back problem one of these years, and try to come to some global understanding of it. Once understood, whether some operator like foi or feu will be necessary to yank us back to “reality” from the subjunctive mood—presumably along with other operators to bring us back from our temporal and spatial forays—or whether some general yank-back will serve, will at last be clear. And we can then decide what to do about Foi or Feu. —JCB
Copyright © 1997 by The Loglan Institute. All rights reserved.