See also Go outside
ISSUE: DEFINITION OF fa'a

fa'a is a spatial tense modal, of selma'o FAhA, which like other FAhA modals describes the location of a referent in space relative to some target, R. Unlike other FAhA modals, that location is not also described relative to a deictic centre (typically the speaker.)

POSITION 1a

The referent is described as situated on a two-dimensional line between some point P and R. (This is in contrast with zo'i, which describes the referent as located on a line between the deictic centre and R.)

Despite its etymology (farna 'direction'), fa'a does not describing the direction or orientation of its referent, but only its location. (Example 3a, not 3b)

PRO:

• Consistency with other FAhA cmavo
• This meaning is the only meaning predicted by the baseline
• Lojban Central has gone on record as thinking it means this

CON:

• Makes fa'a on its own fairly useless
• Doesn't give a ready means of expressing orientation

POSITION 1b

The referent is described as being oriented along a two-dimensional line between some point P and R, with that orientation pointing at R. (This is in contrast with zo'i, which describes the referent as oriented on a line between the deictic centre and R.)

The etymology (farna 'direction') of fa'a supersedes the pattern set by the other members of selma'o FAhA. (Example 3b, not 3a)

PRO:

• TBA

CON:

• TBA

In the case of mo'ifa'a, the referent is described as moving along the line PR. Without mo'i, the condition that the referent is on a line PR is trivial, since P is arbitrary and only R is potentially fixed. (Under position 1b, however, the orientation towards R is itself of interest.)

• How does this square with the consistency position of 1a? mo'ini'a describes a movement towards the location indicated by ni'a. To be consistent, mo'ifa'a should describe a movement towards the location indicated by fa'a, not along some direction which presumably fa'a does not indicate. --mi'e xorxes

• {mo'ifa'a} could indicate that the referent is rotating to a position that is described as {fa'a}. — mi'e pier.

POSITION 2:

It may be useful, however, to draw a distinction between stationary fa'a and to'o: the point P would be closer to the deictic centre than R for the former, and further away for the latter (Example 5).

PRO:

• TBA

CON:

• TBA

POSITION 3a:

When the referent is already known to be moving, the presence of mo'i is orthogonal to that movement. Thus, mo'i is obligatory in specifying the direction of that movement. Without mo'i, fa'a only specifies the location at which the movement is occuring. (Example 6a, 7a, not 6b, 7b)

PRO:

• TBA

CON:

• TBA

POSITION 3b:

When the referent is already known to be moving, the presence of mo'i is redundant to that movement. Thus, mo'i is superfluous in specifying the direction of that movement; this should be done by fa'a alone, consistent with its role as a directional rather than locational particle. (Example 6b, 7b, not 6a, 7a)

PRO:

• TBA

CON:

• TBA

When fa'a is used to qualify a selbri (Examples 1, 2), the referent is the event described by the selbri. The target R is unspecified; P is also unspecified.

When fa'a qualifies a sumti, the target R of the line described becomes the sumti (Examples 3, 4).

Examples

(1) le citno fa'a tavla The young person talks; this happens somewhere between something and something else (NOT: between me and something else, which would be zo'i) (POSSIBLY: The first something is closer to me than the second)

(2) le citno mo'ifa'a tavla The young person talks; this event is moving from something to something else

(3a) le citno cu tavla fa'a le pankaThe young person talks; this happens somewhere between something and the park (POSSIBLY: The 'something' is closer to me than the park: the young person is this side of the park)

(3b) le citno cu tavla fa'a le pankaThe young person talks into the park or The young person talks towards the park (while remaining stationary)

(4) le citno cu tavla mo'ifa'a le pankaThe young person talks; this event is moving from something towards the park

(5) le citno cu tavla to'o le panka The young person talks; this happens somewhere between something and something else (POSSIBLY: The 'something' is further from me than the park: the young person is on the other side of the park.)

(6a) le citno cu klama fa'a le panka The young person is going; this happens between something and the park

(6b) le citno cu klama fa'a le panka The young person is going towards the park

(7a) le citno cu klama mo'ifa'a le panka The young person is going towards the park

(7b) le citno cu klama mo'ifa'a le panka The young person is going; the entire frame of reference this occurs in is in motion towards the park

Archived Discussion

[To be filled in]

Reference Grammar references

[To be filled in]

Usage

[To be filled in]

---

Ensuing discussion raises objections to definitions 1a, 3a, which were the only definitions originally posted. This is an attempt to account for both the (a) and the (b) approaches to fa'a, as an Elephant-like scheme. I (nitcion) haven't finished writing this; please indicate any changes you make to the above clearly.

---

Notice that with this definition, in examples 1, 2, 3, and 5, removing fa'a or to'o leaves the sentence with the same meaning. fa'a adds nothing at all. Even in example 4, fa'a is nothing but a place holder. Apparently a bare mo'i is not grammatical (?!), but mo'i by itself already contains the sense of towards, mo'ini'a is downwards, not moving while down.
Conclusion: with this definition, fa'a is an empty word. No wonder no examples of usage can be found with this meaning.

• Yup; completely agreed. All that is preserved is symmetry with all other FAhA (which, in fact, the entry above is prototyping.) Is the symmetry important enough to preserve for fa'a? Some think yes, some think no. Hence (though I am annoyed about it) I am obligated to write two definitions. And the dictionary will also have to include two definitions for ka, and for vo'a, and for si'o, and for anything there is pending controversy over. ... Which for me is a very, very big minus for the language. — nitcion.

---

Re position 1a:

• Lojban Central has gone on record as thinking that it was 'intended' to mean this; what it actually means is not yet documented, any more than any other usage (or non-usage). I think that the consistent way to deal with orientation, if it were to be a part of tense constructs, would be to add an experimental cmavo (fa'a'a for exemplary purposes) to MOhI. Most orientations that one would want to express probably are covered by the other members of FAhA. Thus we would have:

• ni'a beneath
• mo'ini'a moving downward
• fa'a'ani'a oriented downward

;

Thus we could speak of an event that is beneath the reference, oriented forward, and moving to the left. The baseline grammar would split two MOhI constructs into separate sumti, so the ideal non-baseline solution would be a third optional component of a tense besides space and space_motion for space_orientation, with fa'a'a serving as the tag for this component as MOhI serves for motion. But that is something that could wait for post-baseline to implement, if orientation came to actually be used, given that the current grammar allows a clumsy form with experimental cmavo use in MOhI. (I am not actually proposing fa'a'a, because I'm not convinced that orientation needs to be a part of tense. If Nick or others can identify natural languages that express orientation tense-ly, I would probably feel more sympathy with the idea, since the idea in the tense system was to accomodate as many ways of looking at the world through tense as could be expressed in natural language, as possible, so as to not impose metaphysical bias.) — lojbab

Your criterion that a natural language should include orientation in its tense system, for you to consider fa'a as marking orientation in tense, is bull. We aren't designing Aymara. We're designing Lojban, a language which has gone ahead and made tenses of *everything*, which natural languages don't do. If it is internally consistent for Lojban to also make orientation tenses, then what natlangs happen to do is irrelevant. Lojban shall do what we want it to do, not merely what people have come up with in natlangs so far. Don't impose the metaphysical biases of what has been done already, either. And remember that Lojban has to be internally consistent to work as a language, not merely be a hodge podge of hither and thither linguistic features. — nitcion

---

[This should probably thread off somewhere else]

What the heck? to quote Bob LeChevalier's comments on the mailing list:

Anything not clear in the book may be made clear in the dictionary or textbook when written, but only in response to clear usage supporting the form. The intent for such semantic issues is to "let usage decide".

The dictionary therefore CANNOT rule that fa'a is locational without usage supporting it. Sorry, but the usage opposes it.

In http://groups.yahoo.com/group/lojban/message/8653, the use of fa'a in Alice in Wonderland is discussed as referring to the direction of the event (the 'directional' interpretation) as opposed to the direction to the event (the 'locational' interpretation).
Furthermore, http://groups.yahoo.com/group/lojban/message/8611 gives an example of fa'a being used directionally. During the resulting thread, no-one disagrees and one person comments on my 'good use of fa'a'! The fa'a sentence in question is ".i se lafti le terdi fa'a le tsani" - it rises from the earth toward the heavens.

Lojbab has commanded us to 'let usage decide'. So let's 'let usage decide'. Usage has spoken. The forthcoming dictionary is wrong. And it is not to be published with clarifications not supported by usage. 'Let usage decide' what it already has - fa'a refers to the direction of the event.
- la kreig.daniyl.

1. It is possible for us to decide that usage has decided something incorrectly, and usage henceforth shall be different. This has 'just happened' with ka (countering, I should note, both existing usage and what Lojbab wanted.)
2. "Good use of fa'a" was from xorxes, because he hasn't (hadn't?) changed his mind. That doesn't necessarily reflect a more general understanding of fa'a. And you didn't say anything because...?
3. Lojbab, Pierre, and Nick are on record as agreeing with the above (Nick having changed his mind.) Any Lojban they use henceforth will reflect that understanding. Usage, therefore, hasn't spoken jack; not to us.
4. Inasmuch as the Book says anything, it says that FAhA "specifies the spatial tense of a bridi (the place at which it occurs)". That means a directional interpretation automatically moves outside what the Book describes.
5. If people still think fa'a also or instead means direction rather than location, even though this conflicts with the Book, then by all means feel free to write an alternative definition. The point of the exercise is not what fa'a means, but what a dictionary entry for a cmavo would look like.
6. Normally, I would scour the archives to come up with examples of usage in context (distinct from made up illustrative examples). I think you can guess why I didn't this time.

---

Because the examples disagreed with your opinion?

Because the examples disagreed with my interpretation of what the Book describes as the semantics of FAhA, so yes. (Recall: I changed my mind on this.)

How do you translate the sentence It ascends from the earth up to heaven, and again descends down to the earth, and receives the powers of superiors and inferiors. without wanting a fa'a? The rising toward heaven would get very long-winded. A directional use of fa'a is more often relevant, so while those who disagree with it may be in the majority you don't use fa'a as often. I believe that if we look at the number of uses both ways, fa'a tends to be directional.

('Scuse me, but I translate it as mo'ifa'a, because you're describing motion, not just direction. The 'direction' sense I'm referring to is "speak into the microphone", where there is no motion involved. The direction sense may well be more useful --- but your example ain't it. And it's now up to you to explain why you'd translate your sentence with fa'a instead of mo'ifa'a, or why you'd want this to be fa'a not mo'ifa'a, but motion to the left must still be mo'izu'a not zu'a.)

Using mo'i with a predicate that already by itself describes a motion is in my opinion wrong, because you are then saying that the moving event is itself moving. This could only happen in special situations, for example throwing an object while being inside a moving train, mo'i then describes the movement of the train, the whole event of throwing is moving with it, but mo'i is not the movement contained in renro. To talk about the direction of the throw, mo'i should not be used, so mo'izu'a doesn't work for "throw leftwards". The same applies to lafti. For "throw leftwards" the directional sense of fa'a can be used as fa'azu'a.--mi'e xorxes

• I note the Book has been studiously careful about avoiding double motion situations. I don't like this eliding mo'i business at all, because it makes the semantics of FAhA context-dependent. But I'll have to mull this over a bit. (Dammit, and the only reason I suggested fa'a as a dictionary example is because I thought this was a recent argued but settled debate. The more this goes on, the less I'm inclined to lift a finger to work on this dictionary project. And realistically, I doubt I will anyway...) — nitcion.

• Whoever invented mo'i didn't stop to think for very long before doing it. What was needed was a way to indicate directions based on the FAhA locations, but somehow this was mangled with movement, and the result is that we have two useless cmavo: mo'i that says too much (it adds an unwanted movement to the wanted directionality), and fa'a that doesn't say anything (as a mere location, it is empty of meaning). My preferred solution to this is to ditch mo'i and use fa'a and to'o with their directional sense. — mi'e xorxes

• *sigh* xorxes, any solution that starts ditching cmavo from the language, I have a very hard time justifying entering into the same dictionary that defines mo'i. — nitcion.

• Just mark it as 'evitinda'. [Allusion to Esperanto lexicographic practice, when a form has been used, but is prescriptively judged bogus: "to be avoided". Typically the cultural equivalent of malglico] In any case, if it is used according to its definition it won't be used much. You don't normally want to talk about things such as "eating forwards".

• mo'i was added because we 'did' have examples of motion expressed in tense in natural language. Way back in the beginning, pc and I catalogued all the ways that we thought someone might want to express in tense, using pc's expertise as a tense logician and linguist. We apparently did not think of orientation distinct from position, or perhaps we did and pc felt that we should allow that to be expressed with a subordinate bridi, that being the always-workable alternative to tenses and modals. John Cowan reworked my original more clumsy tense design to be more explainable, and in so doing may have rendered some words like fa'a alone to seem superfluous. But by that time we were averse to removing words or features from the language without a reason that they were a negative feature. — lojbab

• I found it rather clear - mo'i is only elided when the event is already one of moving. After all, we don't say things like lenu muvdu cu muvdu, do we? I have never had cause to describe a motion as being moving outside of hypothetical questions about math or physics. Thus, since mo'i has that meaning, that the event is moving, it makes movement sentences quite confusing. Besides, the direction of the raising in my example is toward heaven. It rises, and the direction of the rising is toward heaven. That is as much direction rather than mo'ifa'a-type movement as speaking with the direction of the speaking being a microphone. The fact that it is moving just obfuscates that, but we want to discuss the direction of movement, not the movement of the movement. - mi'e kreig.daniyl.

• Speaking of the motion of a predicate of moving sounds like the sort of thing that could be very useful in talking about relativity theory and thought experiments. Perhaps someone could try testing this arena of Lojban tense design with a translation/description of one of the thought experiments that went into Einstein's work. — lojbab

• Then there will be yet more usage supporting the idea that "mi muvdu mo'ifa'a le zdani" will mean "I move, and the movement is itself moving toward the house," - and to say "I move toward the house," one uses "mi muvdu fa'a le zdani." - mi'e. kreig.daniyl.

In any case, I will draw up a dictionary entry including both proposed senses — location and direction - and still rejecting with prejudice anything properly covered by mo'ifa'a. This entry will include examples from usage (which I suspect will all go to direction for fa'a), and arguments pro and con both proposals. Hopefully, that will be less noxious to you. nitcion.

• I think that what you've done is quite excellent as a start, though the dictionary entry would have to be grossly abbreviated, with references to the discussion/Elephant rather than the arguments themselves (the arguments do not belong in the dictionary). It might indeed be possible to come up with a way to cover the multiple approaches coherently, once they are fully elaborated. (The only improvement I could request is to find a way to more clearly link the examples to their appropriate definitional segments.) --lojbab

See Elephant, Great Dictionary Problem, The Lojban Dictionary

---

mi'e .aulun. Although not fitting too systematically with FAhA (most of whose members are (just) 'local' indicators) {fa'a} is clearly defined as 'orientation' i.e. static+direction as opposed to mobil+direction (see p. 253 of the Book: "without {mo'i}/with {mo'i}"). There indeed are some more members of FAhA that aren't just static (like {ni'a} and many others) but static+direction like {fa'a}: {to'o}, {zo'i}, {ze'o} what is 'orientation'!
It is a glaring fact and quite obvious, that the Book's (l.c.) misleading when listing e.g. "forward", "backward", ... "upward(ly)", "downward(ly)" etc. in the column '"with mo'i"' and others like "towards", ... "inward (from)" etc. (having the same forms in English!) in the column under '"without mo'i"', but to me it's always been pretty clear and 'natural' to take the static+direction FAhAs as what they are: 'orientation', hence {fa'a} in the sense of "facing/looking at/turned toward".

• le citno cu fa'a tavla (is talking directed to somewhere - not a very good example)
• le citno cu mo'ifa'a tavla (is talking directed to somewhere while moving - not better a phrase)
• le citno cu tavla fa'a mi (is talking toward me/in my direction)
• le citno cu tavla mo'ifa'a mi (is talking while moving towards me)
• le citno cu tavla mo'i mi (is talking while moving to me - not sure whether {mo'i} works 'without' a static or static+direction FAhA attached)
• le citno cu klama mi (is coming to me)
• le citno cu klama fa'a mi (is going somewhere facing me)
• le citno cu klama fa'ani'a mi (is going somewhere - facing downward toward me)
• le citno cu klama do fa'ani'a mi (is going to you facing downward towards me)
• le citno cu klama mo'ini'a mi (is going somewhere - here most probably to me - while moving downward toward me)
• le citno cu klama do mo'ini'a mi (is going to you while moving downward toward me)
• le citno mo'ini'a klama mi (is coming downward to me)
• le citno ni'a klama mi (is coming downward to me) - think it'll work without {mo'i} No! B.S. It's: (beneath the speaker) coming to me.
• le citno mo'ini'a klama do (is going - downward from me/the speaker - to you)

A "verb" already expressing motion doesn't need an additional movement indicator {mo'i}, what would be tautology. But there are those too that have already built-in motion+direction (e.g. "rise", "fall" etc.) and for that reason don't need {mo'i} nor {fa'a}. (Changed my mind: Since {mo'i} does exist in Lojban, it has to be used for unambiguity's sake!)

• Rise (se lafti "be lifted" is what I used in my Emerald Tablet translation) does not have a place for the thing being risen to. I have actually taken to thinking of fa'a as if it were a BAI (adding a place for in the direction of), which works just as well and shows how silly our numerous selma'o are. - mi'e. kreig.daniyl.
  • If fa'a were a BAI, you couldn't combine it with other directions, which shows just how silly complaining about selma'o is. --rab.spir
    • When was the last time you saw fa'a combined with anything but mo'i?
      • When was the last time you saw BAI combined with mo'i? Of course, in my opinion, you should be able to combine {fa'a} with any direction, with the result meaning toward that direction; "moving toward" should be {mo'izo'i}.
        • I process mo'ifa'a separately from bare fa'a. But if my way of doing it works, who are you to criticize? I was only suggesting it in case anyone else would find it helpful. (And complaining about selma'o, at least about how many there are, is not limited to me).

BTW, it's not the {fa'a} but the {mo'i} issue that seems kind of "luxury" to me: from all the natlangs I'm familiar with none is indicating motion like Lojban does. Romanian (just one for all) only indicates direction leaving it to context (i.e. the verb used) whether it's meant static or motional:
de - from, de pe - from on, din (de+in) - from in, despre (de+spre) - from towards, de la - from at, dintre (de+�ntre) - from between etc. There's no difference in form saying e.g. "S�ntem de la Bucuresti" (I'm from Bucarest) and "Am venit de la B." (I came from B.).

• Not an especially nice luxury. Various languages, of course, have words to distinguish orientation from location ("down" vs. "downward"), but Lojban has none unless you use fa'a or fa'a'a for that purpose, because of the amazing screw-up that is mo'i. Apparently, when it was made, nobody was thinking of any examples besides the child walking on the ice... --rab.spir

• Eek. You know, I'm just about ready to capitulate on this, but whatever happens, if the Book has ever claimed a static fa'a (and even if it hasn't, Lojbab and Nora have), that sense should still be recorded in a dictionary, with a comment on how prevalent (or not) it is. — nitcion.

I can't even tell what the sides are anymore. Where's the suggestion I originally saw that fa'ani'a should mean downward? This seems to be what Lojbab is supporting, except with his baselinically correct fa'a'a. --rab.spir

---

Imported from another discussion

There is a gap between simple FAhA cmavo representing nothing but position, and mo'i representing that some arbitrary object in the relationship is moving, with nothing to express simple directions. Since fa'a in its most literal interpretation from the cmavo list (a position that is "toward" something) is entirely useless, it can be reused so that combining it with another FAhA means that the event is oriented "toward" that direction.

Lojbab has proposed this, but using fa'a'a to be safe.

to'o, which appears to be intended to be the opposite of fa'a, can indicate that the event is oriented in the opposite direction.

Actual orientations "toward" or "away from" another object can still be expressed with fa'azo'i for "toward", and fa'aze'o (which could imply that there is something else away from the object that the event is oriented toward) or to'ozo'i for "away".

Examples:

• mi fa'abe'a catlu = "I look to the north"
• mi fa'ane'i klama = "I go inside"
• ko catlu to'ozo'i le carmi gusni = "Look away from the bright light".

This usage is far more useful than anything you can make with mo'i. In "mi mo'ibe'a cadzu", am I walking north, or am I walking on ground that moves north (like a moving walkway)?

Does the child walk leftward on the ice, or does the ice get walked on leftward under the child?