(A page from the Loglan web site.)

(From Lognet 96/2. Used with the permission of The Loglan Institute, Inc.)

The Mia System of Subjunctives

By James Cooke Brown

1. Introduction:

This article is an adaptation of a memo I wrote to the Keugru on subjunctives in February 1996. It was called "The Mia System" after its principal operator, mia, a word I had derived from the Loglan metaphor for "mental place", smina sitfa. I had used this metaphor to define a new complex predicate minsia = is a mental/imaginary world/place in which cause ... leads to effect ..., a world that has been at least in part created by thinker ... . This was, in my opinion, the fundamental relationship underlying the subjunctive mood; and minsia was therefore at the heart of the subjunctive system I was designing. Moreover, minsia could be further shrunk into the CVV-form mia, an easily remembered version of the PA-type operator required by the system. However, at the time, the CVV-form /mia/ was not, strictly speaking, available to me, as years before, being a handy contraction of mi ze da, mia had been assigned to the "I and someone else" sense of we/us/our.

From recent talk in the Keugru about the da series in general I expected the /miV/ series to be imminently returned to the CVV pool; and so I felt free to use mia--tentatively, to be sure--as the characteristic operator of my new system. But by the time of this writing, the word is so identified with that system among its partisans that they refuse to use anything else, soi crano. Still, let me state firmly at the outset that my use of it in this paper is for expository purposes only. Any of the other four /miV/ words could be used as the subjunctive operator if the decision is made to retain /mia/ as one of our we's. What I think is extremely likely is that, if the Mia System is adopted, the first two letters of its principal operator will continue to be /mi/, as these are firmly given by the combining form min of smina, the Loglan word for mind or mental. For whatever else the Mia System is, the subjunctive mood as interpreted by it is a strictly mental operation that has little or nothing to do with the external--sometimes incorrectly called the "real"--world (as if mental worlds were not).

Allusions will be made in this article to the two other subjunctive systems the Keugru has been studying. The earliest to be put before us was the "Foi/Fio System", the one presented by James Jennings in his article "I Would If I Could" in Lognet 96/1. James's article was based on a proposal made principally by Emerson Mitchell, but joined by our other logician, Randall Holmes, and contributed to by two other members of the Keugru as well as by James himself. Despite this doughty provenance I devised a second system after studying the Foi/Fio one, and I called it the "Sio/Dau/Biu System" after its principal operators...more frequently the "S/D/B" or "Two-Slot" one. S/D/B was designed to fill in the linguistical lacunae I believed I had found in Foi/Fio.

Both these older systems for managing the Loglan subjunctive have been proposed to the Keugru in the course of the last year-and-a-half, the period during which we have been conducting this intriguing inquiry. Foi/Fio had actually emerged from Keugru discussions in what is practically its present form by May of 1995. S/B/D was developed in response to it in the summer of 1995 but has never been described in hardcopy. Probably it never will be, though an e-mail memo presenting it was made available, first to the Keugru in July 1995, and then to logli who asked for it in September of that year. The reason I say that the details of S/D/B probably never will be broadcast is that I, its author, have since abandoned it in favor of the more mature Mia System, this newer model of my thinking having incorporated most of the insights of S/D/B as well as the major feature of the F/F system, namely its treatment of conditional relations. In short, the system I am about to describe owes much to both earlier systems.

The article that follows is an only slightly modified version--mostly augmented by this introduction and by a fresh anecdote or two--of my February 1996 memo to the Keugru. Bear in mind that it was originally intended for my forbearing colleagues in that talk-tolerant body, soi crano, so it is written in my "leave no stone unturned" expository style. I have not tried to remove all traces of its origin in that collegial context.

2. A New Subjunctive System:

Here's a third subjunctive system. I'm calling it the "Mia System", for reasons that will emerge. In some respects it's like Foi/Fio; in others, like S/D/B. But it's not a synthesis. I believe it embodies a wholly new insight into what subjunctives actually do for natural language speakers; and I have come to believe that we should satisfy that function with our new subjunctive operator, and no others. In other words, I believe we should disentangle subjunctivity from the several other linguistic operations with which it is frequently confused.

Therefore the Mia system, unlike the S/D/B one, which was two-dimensional, is a three-dimensional system in which (a) subjunctivity, (b) effect-liklihood, and (c) factuality, or "world-liklihood" are all treated independently of one another. Let's consider its primary dimension, subjunctivity, first.

3. Subjunctivity with Mia:

Consider a predicate with the meaning:

1) X is an imaginary/conceivable/"possible" world in which event/state Y occurs/obtains and causes--"would cause" if its world existed--event/state Z, and this world and its causal laws have been at least in part imagined by thinker W.

This predicate has, I believe, all the ingredients of the natural-language subjunctive. It allows the speaker to claim that there "exists"--simply because he can imagine it--a world X that is characterized by a certain early event/state Y that would then create/make possible/provide the conditions for--in a word, cause--some later event/state Z in which he has some interest, and it relates that world to its creator W, who is usually the speaker him/her-self. Thus, using this predicate I could say:

2) If I were king (event Y) I would, in the imaginary world (X) in which I was king, be happy (event Z).

Notice that we need not mention the world-creator W because, unless we do mention it, convention permits the listener/reader to assume that W is the speaker/writer of the embedding sentence.

This predicate handles everything that is in any sense subjunctive, including mundane "profactual" remarks (that appear to be about the "real"--that is, external--world but are not):

3) If you'd bring me that chair (Y), I could reach this lightbulb (Z).

Here the imaginary world X is a certain minor rearrangement of the furniture. But it's a rearrangement that has been carried out first in the speaker's mind, though perhaps later that arrangement will be replicated on the ground. No direct allusion to the imaginary world X is made in (3) but it's surely there. As we might say, the new furniture arrangement "has already been made in the speaker's mind".

Another profactual example--"profactual" in the sense that it, too, like the world of (3), can easily be realized in the external world--is:

4) If a tornado came to my town (Y), I would seek shelter in my cellar (Z)

Again, the world X, in which cause Y and effect Z are being linked, has been created in the mind of W (who is also the speaker); but nature can also create this world, and when and if nature does create X, and if W knows himself reasonably well, we can expect cause Y to be linked to effect Z in it as well.

Here's a counterfactual example...one not realizable in the external world under any circumstances--not in the external worlds of most people, anyway--but just as easily imagined in the internal world as the shifting around of furniture is, or one's own behavior in emergencies:

5) If your dead mother were alive today (Y), she would tell you to find a nice girl and marry her (Z).

Again, the imaginary--and impossible--world X in which Y and Z would then take place has been invented in W's mind. But this time no possible rearrangement of the external world can get to it. It is, as we say, "counterfactual". Moreover, this is, I think, all we mean by the English word counterfactual; it means you can't get to it by any pathway in the external world.

But the point is that the very soul of the subjunctive is not its counter- or profactuality as I once thought--in this respect I have come to agree with the early partisans of the Foi/Fio system--nor is it the liklihood of the effect's occurring in the imaginary world, which is the would/could distinction made by the designers of the Foi/Fio system, as well as one of the "slots" in my own S/B/D one. In this, I now disagree with both my earlier self and with the present partisans of F/F. Instead, I now believe that what makes any sentence a subjunctive one is its report of the creation of some imaginary world in the mind of some imaginer W, who is usually the speaker, whether W additionally believes that he/she can reach that mental world by traveling in the external one, or not.

We may also note that, invariably in the subjunctive remarks made in natural languages, causes and their effects are reported to take place in the imaginary worlds created by their speakers in perfectly regular, knowable ways...even though they are taking place only mentally, and sometimes inaccessibly. This is the remarkable thing about these "subjunctive places" that seem to be spontaneously generated in the course of human speech (and no doubt, too, in the speech of chimpanzees and gorillas, had speech evolved among these animals, too; for it is now well-known that they too are evidently "insight"--that is, imagination--driven in at least some of their problem solving). Apart from the special features the imaginer has put into them--once-dead mothers counseling sons, or commoners becoming kings--many features are just like those of the external worlds we live in.

From this point of view, then, the subjunctive mood is simply a linguistic device for alerting auditors to the fact that some speaker has entered an imaginary place in his/her mind, and is talking about the events and their possible causes taking place there. On this view, the reason we find subjunctive moods in so many natural languages is because we humans are very prone to do this. We are prone to invent, and once invented to talk about, imaginary worlds and the often frolicsome things taking place in them.

That, in a nutshell, is the theory of the natural language subjunctive that underlies the present essay, as well as providing us, as it does, with the eliminative predicate and the PA-type operator that logic requires of the Mia System.

Before turning to the linguistic nuts and bolts of this system, however, permit me to share with you two anecdotes that show the English subjunctive in action. I tell these stories in the hope that they will persuade the reader that, in viewing the subjunctive as a tool of the human imagination, we are probably on the right linguistic track.

For the last year-and-a-half, ever since our logicians began to talk to us in the Keugru about "counterfactual conditionals"--as well as about other logically interesting manifestations of the subjunctive mood--I have been observing the use of the subjunctive in the speech and writing of plain people. For instance, while traveling in Australia this Spring my wife and I heard an Irish comedian tell the following joke:

Pat says to Mike, "Have you heard that Sean O'Casey died?"

"No!" says Mike. "When'd he pass away, then?"

Pat rubs his chin and says: "Well, if he'd a lived 'til next Thursday, he'd have been dead a month!"

I don't think I've ever encountered a clearer instance of a speaker inviting a listener to join him in a private world...in this case, a preposterous one in which Sean O'Casey is visibly both dead and alive. Thus in Pat's subjunctive world, the Law of Excluded Middle is itself set aside. His friend Sean has to be both dead and not dead for this convenient calculation to work!

And yet it does work! That is one of the most elegant things about our use of subjunctive worlds. We can evidently rearrange the "realities" of our external worlds so cleverly as to be able to explore all sorts of fancies in them...sometimes quite useful fancies, such as what one is going to do in an emergency.

Here's another Aussie use of the English subjunctive. This one is quite folksy, less dependent on the faerie worlds of the Irish imagination, but equally insight-provoking even so:

A tourist lady is rejoining our tour bus and glances smilingly at another tourist lady, already seated, as she passes down the aisle. The seated tourist lady says "Oh! Have I taken your seat?"

"No," says the smiling tourist lady; "And if you had, it wouldn't have mattered!"

Again, we have a speaker displaying publicly a feature of a privately imagined world. The amazing thing is that we all trust the speaker of this graceful sentence to "know" the causal laws by which she runs events in her own imaginary world, and so we trust her report of this causal linkage! We trust her, in short, to know herself so well that she can speak "truly" of an imaginary situation in which she has probably never found herself before!

What an extraordinary linguistic invention the subjunctive is! I'm afraid I have grown quite fond of it during these months that I've been studying it. Like an entomologist studying wood-lice, I grow fonder of my little subjunctive critters the more I watch them.

4. The Basic Mia Plan:

We can mobilize the Mia plan by making a complex preda with the meaning of Section 2's specimen (1) and deriving a CVV-form operator from it. I rather favor the metaphor "mental place" for these imaginary worlds. We may have to use the more straightforward metaphor "mental world" for this notion, after constructing a new, and this time primitive predicate for the notion of "world"; and then use one of its combining forms to replace the last half of /minsia/, presumably one not ending in /a/. But as these lexical matters aren't settled yet, I will in the rest of this essay use minsia for the as yet unknown complex predicate for is a mental/imaginary world, etc.; and mia for the operator to be derived from it.

With these two tools provisionally in hand, then, we can now translate such classical English subjunctive claims as:

1) If I were king, I would be happy.

into straightforward if novel Loglan:

2) Ba minsia lepo mi bragai guo, lepo mi hapci.
Something-x is-an-imaginary-world in-which my being-king (end clause) (causes) me to-be-happy.

This is the fully detailed claim behind (1). (Well, not quite fully detailed, of course, because we have accepted by default that the builder of this "something x" is the speaker.) And now we can express (2) in briefer compass with the "subjunctive operator" mia:

3) Mi hapci mia lepo mi bragai
I am happy in-an-imaginary-world-in-which I am king.
I would/could be happy in a world where I were king.
I would/could be happy if I were king.

We notice that this is approximately the same way that subjunctive conditionals are expressed in the Foi/Fio system: a PA operator is used to attach the condition, or cause, to a main sentence that asserts the effect. The main difference is that foi/fio make the would/could distinction and mia does not. In short, we are borrowing the major insight of the earlier Foi/Fio system to build a new system which, we hope, will be more truly subjunctive from a linguistical--that is, natural language--point of view.

Notice what we are doing in (3). We are treating the cause of interest (mi bragai) as a condition; and we are expressing that condition in a LEPO-clause as the argument of a PA modifier. Moreover, the whole phrase so generated is a "sentence modifier" in the L1 sense that it is free to be moved anywhere we want to put it. Meanwhile, the effect of interest (Mi hapci) is stated as the main claim of the sentence: the claim modified by the PA-phrase, for its conditionality is now safely managed by that phrase. This maneuver is, of course, the major achievement of the Foi/Fio system, which was itself apparently a borrowing from Kripke's "modal logic". I have happily retained it.

But now notice two other things about Mia. It has two features that differ from those of earlier proposals: (a) as mentioned earlier, the distinction between foi and fio has been dissolved; mia is apparently to be used with all kinds of subjunctive statements, the more confident "would" kind as well as the less confident "could" kind; and (b) a major claim, namely that there is an imaginary world in which this cause-effect relation obtains, is now being implicitly asserted in the compressed fashion afforded by the PA operator. For just as pa itself permits the compression of statements that can only be fully made with pasko, so mia permits the compression of statements only fully made with minsia; and minsia explicitly claims that the world in which these events are taking place, and in which this causal linkage obtains, is an imaginary one. Evidently mia is a blend of foi and fio and presupposes--as they don't--the existence of an imaginary world. With mia, in short, we break out of the domain of the "logically possible". For we are now as free as Pat was to entertain logically impossible worlds in which dead friends are both dead and alive!

This makes of (3) a very modest claim indeed, nothing like so strong as claims made with the F/F system with its modal logic. For all that is required for (2)--which is the formal interpretation of (3)--to be true is that the speaker be able to convince us that W has actually imagined a world in which such things take place, and that he/she is that W. So this more modest subjunctive system allows any speaker s to make a bare-bones claim that some world X exists--presumably in s's own head--in which this causal relation between Y and Z obtains.

Are such claims then simply too empty of factual content to be quarreled with? No; the auditor of (3) can retort But I know you better than you know yourself! You wouldn't be happy as king! Putting only the last of these counterclaims into the new L:

4) Tu no hapci mia lepo tu bragai.
You would not be happy if you were king.

Eliminating mia with minsia, as above, we get the following interpretation of the auditor's counterclaim:

5) Be minsia lepo tu bragai guo, lepo tu no hapci.
There is an imaginary world y--namely the one in my head--in which your being king causes you to be unhappy.

There is a quarrel here, alright, a genuine difference of opinion; but it's a weak one.

The Mia system can also make this quarrel an honest one. The Loglan interpretations of the two claims reveal that two quite different imaginary worlds have been imagined by the two claimants. They will discover, as logli, that on careful interpretation of their remarks (5) is clearly not a refutation of (2); both could be true and probably are. So with the Mia system in place, Loglan could help people resolve such "subjunctive disputes" amiably...which is exactly what a logical language ought to be able to do for them. What logli will often discover is that they are talking about two different imaginary worlds...and therefore, perhaps--or so, as a dues-paying optimist, one is perhaps allowed to hope, soi crano--be somewhat less likely to pick up clubs and start bashing in heads. For this is what we humans have unfortunately done all too frequently in the past about just such differences among our imaginary worlds.

That a logical language should be a moderating force in such fallacy-laden disputes, is of course just as it should be. In natural languages, I've observed, people cannot, in truth, sensibly refute one another's subjunctive claims by peaceful means. That is why logicians used to call such "reasoning" fallacious. (I still do.) When I was a boy--and later, when I was a young university teacher of Freshman logic--the fallacy was called "hypothesis contrary to fact". Our calling such reasoning fallacious cautioned our students against such nonsense as (amateur) historians are inclined to generate when they argue with one another about "what would have happened" had the Greeks not defeated the Persians at Thermopylae. Under the Mia system, the fallaciousness of such disputes will be plain to the logli who make such competing claims...and are willing to face what their words actually mean when they do so.

Not all subjunctive arguments are, of course, as empty as the one about Thermopylae; but many are. Most theological arguments, for example, are completely subjunctive. It is only when speakers are able to move their subjunctive predications with them into the external world--the intersubjective world of science that they can both observe--that refutations of one another's once-subjunctive claims can then take place...usually retrospectively. (See? says the king's childhood friend. You're not happy at all, now that you're king!)

But if this system for expressing the subjunctive is adopted, the subjunctive claims of Loglan will, as I am arguing they should be, be very modest claims indeed; and they will usually be quite untestable. Although such claims may seem to contradict one another, actually they never do. For on this model of what is going on, subjunctive claims are about what is taking place privately in different speakers' heads.

Every now and then someone, like the smiling lady, lets us into her private world. And we are often stunned or charmed by what by what we are permitted to see there. But that doesn't make people's claims about them refutable, or even very arguable. Who knows what the smiling lady would really have done had her seat been occupied? Who is best positioned to know about the correlation, or the lack of it, between her private and her public worlds?

So we believe her. Who else could know?

5. Missing Conditions:

What about missing conditions, then? Subjunctive remarks in which the condition, or cause, is left unspecified are perhaps the most frequently occurring uses of the subjunctive in natural languages. In the Mia system, such "incomplete subjunctives" will be easy to express. The speaker will just say mia ba to mark his claim as a subjunctive one without bothering to specify what ba is. So when a Loglan speaker s doesn't wish to specify the causal element in s's imaginary world--the one that is bringing about the effect s is talking about--s can say:

1) Mi barteu ta mia ba.
I reach that given-some-unspecified-cause x (in some unmentioned imaginary world y).
I could/would reach that.

To which a helpful response might be: Hu nu cnida tu? (What do you need ?) Or Hu nu cnida tu lepo barteu da? (What do you need to reach it?) Both of these responses are based on an understanding that hidden in ba is an already imagined tool.

Another example:

2) Mi hapci mia ba.
I am-happy given-some-unspecified cause x (in some unmentioned imaginary world y).
I could/would be happy.

The invited response now is, Ba bi hu? (Under what imagined circumstances?). Or

3) Mi hargoi le hasnilca mia ba.
I seek-shelter-in the cellar given-some-unspecified-imagined-cause x (in some unmention imaginary world y).
I would/could seek shelter in the cellar.

Again the logli auditor is invited to say, Ba bi hu?

In sentences (1-3) it will have become clear why the Mia system makes no distinction between English would and could. This distinction seems relevant in some subjunctive contexts but not in others; and we will deal presently with how to make it explicit when it is relevant. Obviously the would/could difference is a nuance...a real one, but a nuance even so; and it is, I believe, wrong to turn it into an obligatory feature of the subjunctive system, as was done in both the Fio/Foi system and the Sio/Dau/Biu one. Whether would or could is the correct word in an English translation of a Loglan subjunctive is often quite irrelevant to the subjunctivity of the remark, which is fully conveyed in Loglan by the marker phrase mia ba.

But in general, I'd recommend using the English word would to translate mia ba in these "bare bones" subjunctives, leaving the distinctions between could, might, should, must, etc. for the several optional embellishments of mia sentences that I'll develop later.

6. Effect Liklihood:

The differences between would (certainly)/should (probably)/might (possibly) are all differences in what I'm going to call "effect liklihood", and these differences were expressed in the S/B/D system by the operator triplet sio/dau/biu that gave it its name. (In F/F, only two of these probability grades are present: would (certainly) and might/could (possibly).)

Every subjunctive remark specifies some effect of interest: happiness, reaching the lightbulb, seeking shelter in the cellar, one's seat being occupied not mattering, being dead a month, etc. And that effect can be regarded as having a liklihood that is strictly conditional on whatever cause is supposed by the speaker to bring it about. If any of these conditional probabilities is attached to a subjunctive remark, then both cause and effect are believed to be capable of happening in the imaginary world that the speaker has devised; but now the "firmness" of the imagined causal connection is also being specified.

Sometimes it is important to the speaker to specify the liklihood of such imaginary causal linkages; sometimes it is not. In the Mia system, specifying event-liklihood is strictly optional. In both the F/F and S/D/B systems it was not One could not speak subjunctively in either of these earlier systems without mentioning the liklihood of the effect of interest. It was as if, to the designers of these earlier systems--which includes an earlier version of myself--effect-liklihood was to be an obligatory feature of the grammar of the subjunctive, just as tense is an obligatory feature of the grammar of the English verb. But as Whorf has taught us--and as I remembered in time to incorporate in the design of Mia--obligatory grammatical arrangements are rich sources of the towering "metaphysical burdens" we speakers of natural languages unconsciously carry around with us inside our heads. As the kerju in the Loglan Keugru, we Academists are committed to lightening that burden, or to keeping it light where Loglan already makes it small. So if we still mean to perform this Whorfian service for future logli, we must continue to avoid obligatory grammatical features wherever we can. With Mia, we can mention effect-liklihood whenever we wish; but we can also choose not to make these probability assessments of events in our imaginary worlds. So in addition to using mia as a sign of subjunctive intent, a logla speaker may freely choose to specify effect-liklihood alongside it, or not.

The tools used in the Mia system to express effect liklihood are, as I indicated above, the three probability operators of the earlier S/D/B system. In building these operators into the Mia system, I am borrowing from my own earlier work. (Those among my present readers who are e-connected and interested in how that work took shape can ask to see my "Two-Slot" memo of last fall.)

The three probability operators sio/dau/biu were, unsurprisingly, central to the S/D/B system. But even if they are not used to express the subjunctive any longer, they are still members of the PA Lexeme, two of them old (dau and biu), the other (sio) new. Old or new, they were made by deriving them from a trio of predas: sirto/dakli/blicu, sirto having been suitably redefined for the occasion and the definitions of the other two being regularized to match. The three operators themselves thus became the objective versions of ia/io/ii in that, with them, one can now make probability claims about other claims. Let us see how.

Suppose we are zoologists and have found a specimen of a previously unknown animal. We examine it and say:

1) Da sio mamla.
X must be/certainly is a mammal.

Or we may decide:

2) Da dau mamla.
X should be/probably is a mammal.

Or we may say:

3) Da biu mamla.
X might be/possibly is a mammal.

All of these are hard, testable, intersubjective claims about claims. They may be interpreted, as all claims involving PA inflectors may be, as:

4) Lepo da mamla ga sirto/dakli/blicu ti ba.
The-state-of X's being-a-mammal is certain/-probable/possible given-these-conditions (those presumably present in the world of speech) and-some-unspecified-knowledge-system-x.

for that is how sirto/dakli/blicu are now all defined. This can lead to close scientific questioning about just what "these conditions" are and what that "knowledge system" is...especially if we're being taxonomically fussy, as occasionally we should be, as scientists. But normally we non-zoologists just take such sio/dau/biu embellishments as simply "objective" versions of ia/io/ii. We know, when we hear them, that their speakers believe their remarks to be solidly based on taxonomic fact; that they are not just expressing "feelings", but giving "considered opinions" about those facts. But note that as expressions of attitude, the ia/io/ii operators have not been disemployed by sio/dau/biu. They still work admirably to display our attitudes toward what we say:

5) Ia/Io/Ii da mamla.
Certainly/Probably/Perhaps X is a mammal.

Here, all three claims are simply Da mamla. The attitude indicators are just that: indicators of the speaker's attitudes toward s's own claim. So sentences (5) are not claims about claims as sentences (1-3) are; they are claims about the world concerning which the speaker happens to have some "epistemic" feelings.

Incidentally, to distinguish between sio/dau/biu and ia/io/ii--as I am now proposing that we do quite strictly--is only to take yet one more step toward one of Loglan's most ancient goals: to make the distinction between the emotive content of speech and its cognitive content as plain as possible. Adding sio/dau/biu to the language will help us do that...quite independently of how or whether we use them as optional embellishments of the Mia System.

But to return to the burning question, soi crano: How can we use sio/dau/biu as "optional embellishments" of the Mia system?

We can develop an answer to this question fairly quickly by showing, first, how we can eliminate these three operators whenever we do use them. That will surely suggest ways to use them, soi crano.

Obviously we can eliminate sio/dau/biu with the three probability predas sirto/dakli/blicu, as in (4) above. Each of these three predas apparently has the form:

6) X is certain/probable/possible under conditions Y given knowledge system Z.

In a "Mia sentence"--one embedding a mia phrase--X is the effect of interest: Mi hapci; Mi barteu le litbui; Mi hargoi le hasnilca; etc.; and the operand of mia is the cause or condition Y: ba; lepo mi bragai; lepo tu berti dio mi leva cersi; lepo ne tontetri ga godzi levi sitci; etc. The mia operator turns this effect X into one of "subjunctive interest" by announcing the existence of an imaginary world in which that effect, say happiness, will, with some liklihood, take place as a consequence of some prior cause or condition Y, say kingship, that is already taking place in it.

But what we want to do now is specify the degree of that liklihood: the probability of that effect's happening given that cause, as these are thought by the speaker to be related in this imaginary world. We can express this as the liklihood that the effect of interest X will happen in that imaginary world given the prior happening of cause Y in it. We want to say whether, given Y, X is certain to happen in that world (P = 1), only probable in it (P > 0.5), or, the lowest degree of all, only possible in it (P > 0).

Certainty, probability, and possiblity--augmented by their negatives--are the three degrees of probability that natural languages seem regularly to provide their speakers, whether they are talking about subjunctive worlds or external ones. When applied to subjunctive worlds, there seems to be a "natural" (hard-wired?) partitioning of words into three probability-grades, e.g., English would/should/might. We will run into triplets of this sort again.

Let us apply this reasoning to our classic Mia sentence: specimen (3) of Sec. 4, which we copy here:

7) Mi hapci mia lepo mi bragai
I would/should/could be happy if I were king.

By convention, lepo mi hapci designates the effect of interest; lepo mi bragai, the cause; and the imaginary world in which the causal relation between it and the effect obtains presumably lurks in the background. How to apply sio/dau/biu to this curious situation? It turns out to be quite easy.

Taking our cue from one of the results of the S/D/B study (which I don't have space to go into here), we see that

8) Lepo mi hapci ga sirto/dakli/blicu lepo mi bragai guo ba.
My being happy is certain/probable/possible given that I am king and that a certain unspecified knowledge system x exists.

We can now express these same relations more economically--pushing the knowledge system, too, into the presumable background--as:

9) Mi hapci sio/dau/biu lepo mi bragai.
That I am happy is certain/likely/possible given that I am king.

Notice that (9) says nothing whatever about imaginary worlds; for (8-9) are not subjunctive remarks. They are indicative remarks about causal relations, presumably ones obtaining in the external--i.e., non-mental--world. The only thing that is unspecified in them--because they are abbreviations--is the knowledge system by which such probability calculations can be made.

We now notice that, except for the prepositional operators, sentences (7) and (9) are exactly the same sentence. This invites us to combine those operators by "compounding" them, just as we combine /pa na fa/ to get the compound tenses papa/napa/fapa, etc. We call the words we will now create the "compound subjunctive operators"; and we'll form them by joining each of /sio dau biu/ to /mia/:

10) Mi hapci siomia lepo mi bragai.
I would (certainly) be happy if I were king.
11) Mi hapci daumia lepo mi bragai.
I should (probably) be happy if I were king.
12) Mi hapci biumia lepo mi bragai.
I might (possibly) be happy if I were king.

Now these are genuine metalinguistical claims. They may be subjunctive ones, but they are claims. The fact that they are subjunctive claims and not indicative ones is told by the -mia part of each compound. But each claim now presupposes not only the existence of an imaginary world but a knowledge system capable of telling us how to calculate the probabilities of causal relations in it! So the compound claims are simply finer-grained than the simple ones made with mia are. They claim substantially more about their subjunctive worlds than mia alone does. These three prefixes, then, add the "effect- liklihood" dimension to the Mia system. They "fatten" the subjunctive world by giving it this optional second dimension.

I leave the reader to figure out the eliminative formulas for these compound subjunctives. The main thing to remember in developing them is that there are now two implicit elements to be made explicit in the interpreting claim.

7. "World-Liklihood":

When I first began to think about the issue of "factuality" in the Mia system--i.e., whether the imaginary world presupposed by mia is "accessible" (as our logicians call it) from the external world or not, and if it is, just how accessible it is--I thought that speakers' beliefs about this important dimension of subjunctive usages could also be succinctly expressed, and by the same three probability operators that are used to express effect-liklihood, but attached to "something else".

But I see now that that is not possible. The "factuality" relation obtains between worlds, not between causes and effects. It is a relation between the imaginary world "ba", whose existence in the speaker's head is implicitly announced by mia, and the external world ti, which s is likely (by another probably hard-wired mechanism in the human brain: the "mind-body" separation mechanism) to think of as the only "real" world. In fact, of course, the external world isn't any more real than the worlds generated inside our heads. But setting aside the knotty neuroscience question of what sorts of "reality" are generated inside human heads, we can represent by ti the world existing outside it at the time of speech, just as we can use ba to represent the imaginary world s has created in it. Thus, what we are looking for is an elegant way of expressing the "factuality" relationship between ba and ti.

But no representation of ba, the imaginary world, is present in the mia sentence! Only the effect of interest (Mi hapci, for example) and its cause (lepo mi bragai) have any linguistic representation in this type of sentence. And these two causally-related events have nothing whatever to do with the "factuality" of the unmentioned imaginary world ba.

Now all members of the PA Lexeme relate the event of the main sentence to something or other. So factuality cannot be reported by mia--or by sio/dau/biu, for that matter--in any well-defined position in the Mia sentence. Given this constraint on their use, we must abandon any hope of using PA operators to express the factuality dimension of a subjunctive sentence.

However, we can take the cause claimed to be operating in the imaginary world ba as a "representative characteristic" of it, and designate it by that means. We may then say that the "accessibility" of the world so characterized from the world of speech (which we can always designate with ti)--that is to say, the probability of the one world developing into the other given all we know--can be mapped onto the same probability scale that we used for effect-liklihood.

Let's try this trick and see what we get. Interestingly enough, it will turn out that this is exactly what we do in E when we sadly mutter but I'll never be one after saying wistfully I would be happy if I were king.

When English speakers wish to specify the pro- or counterfactuality of their subjunctive worlds--that is, the world-liklihood of their subjunctive remarks--they append such supplementary clauses as but it'll never happen or which is impossible or and that's quite possible, and so on.

If we can accept this as a compact way of talking about the accessiblity relations between worlds in Loglan as well, then the problem is simply to find a way of executing such embellishments neatly and compactly.

For starters, let's take

1) Mi hapci mia lepo mi bragai.
I would be happy if I were king.

How do we add the English mumble ...but I'll never be one? Actually, quite simply. By using our alphabetic anaphora, we can co-designate lepo mi bragai by the letter word bei (the first letter of its principal preda) and say

2) Mi hapci mia lepo mi bragai, ize bei no blicu.
I would be happy if I were king, and-jointly b (the event of my being king) is not possible.

Notice that these are joint sentences, not merely connected ones. The events they claim separately must happen more or less at the same time and place for the claim they jointly make to be true. (In other words, this is a set of claims, not a multiple.)

Now we can take such subjunctive-supplementing remarks to mean, not only that the speaker believes that the event of s's being king is impossible in the real world, but that the world in which s is king is "inaccessible" from the world of speech, in exactly the sense given that term by modal logic.

Here are some other variations on this "world-liklihood" theme:

3) Mi hapci mia lepo mi bragai, ize bei blicu ce no dakli.
I would be happy if I were king, and-jointly b (the event of my being king) is possible but not likely.
4) Mi hapci mia lepo mi bragai, ize bei dakli ce no sirto.
I would be happy if I were king, and-jointly b (the event of my being king) is likely but not certain.
5) Mi hapci mia lepo mi bragai, ize bei sirto.
I would be happy if I were king, and-jointly b (the event of my being king) is certain.

So any degree of accessibility between the world of speech and the imaginary one signalled by the use of the subjunctive can be specified in the Mia system, and in reasonably economical ways.

This solution to the problem of conveying the speaker's beliefs about the pro- or counterfactuality of s's subjunctive remarks does not have the brevity that we ordinarily like to achieve in Loglan usages ...especially when we expect or want them to be widely used. But the convention I'm proposing does say exactly what is required with reasonable compactness; and it is probable that the world-liklihood dimension of subjunctive usage is the least commonly specified of the three dimensions we've considered.

So Zipf as well as Whorf would likely be satisfied with these arrangements, soi crano...to speak subjunctively for a change.

8. Five Other "Auxiliary Verbs":

When and if we provide Loglan with a subjunctive mood, and our system for doing so turns out to be the one described above, or one similar to it, we might then use the sirto-sio-ia triplet that the Mia system employs as a kind of "template" for creating other Loglan "auxiliary verbs". Using this lexical mechanism we may be able to express these other "auxiliary" relations as economically and objectively as we will then be doing in the "other parts" of the subjunctive system. But these so-called auxiliaries are "parts" of the subjunctive system only in the sense that, in natural languages, they are often confused with it. The words I have in mind here are such English auxiliaries as can/could, will/would, and may/might, should, must in their non-subjunctive senses.

What is required in each of these five cases is a predicate word that can express the objective sense of the required relation. That word can then be used to build a CVV-form allolex of PA that will contrast with the attitudinal sense we already have for each of these five notions except can. Thus ai/oi/oe/oa are already in the language as attitudinals. We are about to build their objective equivalents...as well as a fifth auxiliary for can (which has no attitudinal equivalent).

What is required is five sirto-type predicates, each of which can involve, as one of its places, the specification of the conditions under which something can, may, must, should, or be intended to happen. Here's a set of predicates that would satisfy this condition. Only one of them, ?nurlei, is new to us. The rest are taken from LOD.

From these five predas we can build fuo/kau/fui/foi/nui as the required set of "auxiliary operators". Currently unused CVV forms are available for all five of these new words, though /foi/ is now being reserved for the Foi/Fio system in case that should be the one adopted. With that proviso, then, we can now use these words as we would any other PA word, that is, in inflecting, adverbial, or prepositional positions. Take fuo = intend to from furmoi as our examplar:

1) Da fuo godzi la Espanias.
X intends to/will go to Spain.

Attitudinally, one can still use ai to say Mi ai godzi la Espanias, when it is a claim about one's own behavior one is modifying. Or one can use ai to express one's intent to force someone else to do something, for example, Tu ai godzi la Espanias; but one cannot use an attitude indicator to "express" another person's intent! So (1) gives us a handy, abbreviated way of claiming to know something about another actor's intentions.

(1) may be interpreted as

2) Da furmoi lepo da godzi la Espanias, guo ba ti.
X intends-to-bring-about the-event-of X's going to Spain, for some unspecified reason x under "these" conditions (i.e., the conditions that obtain at the time of speech).

We can also express this same claim adverbially:

3) Da godzi la Espanias fuo.
X goes to Spain intentionally.

and a useful form results when we use fuo prepositionally:

4) Da godzi la Espanias fuo tio.
X (intends to) go-to Spain under those conditions (i.e., under the conditions last mentioned).

A deferred auxiliary seems very odd to our English-speaking minds; but perhaps we can get used them...as we'll have to, of course, when the adventurous among us begin to use them in this way!

Kau = can, fui = should, nui = may, and perhaps foi = must if available, may all be used in similar ways with similar interpretations.

The importance for English-speakers of adding these five auxiliaries to the language at the same time that we add a subjunctive system to it is that they will facilitate sorting out, into precise Loglan, the many ambiguities and vagueries that abound in these regions of the natural languages we speak. It is especially important that we English-speakers be able to do this; for it is by these overlapping and mutually obscuring forms of English that our minds have probably been confused (I hope not unalterably!) about the subjunctive mood! By disentangling these linguistic themes, then, I hope we can deconfuse ourselves on the matter of the subjunctive and its strange linguistic bedfellows.

. . .

I feel I've given you a lot to think about, perhaps too much for a single go. But then there is "too much" to think about! When we dared to even think about adding a subjunctive to our logical language, we opened a very large can of worms.

But then, "somebody's got to tame them", soi crano. What else are logli for?

Copyright © 1996 by The Loglan Institute, Inc. All rights reserved.

Send comments and corrections to:

djeimz AT megaseattle DOT com