View source for The proposed fourth tense of Lojban

Originally stored at [http://www.lojban.org/files/papers/4thtense www.lojban.org/files/papers/4thtense].
<pre>
Subject:  The proposed fourth tense of Lojban
Participants:
	Arthur Protin
	Bob LeChevalier
	Carl Burke
	Doug Landauer
	Guy Steele
	Jack Waugh
	Jeff Prothero
	Jim Carter
	Robert Chassell


[1] by Jeff Prothero

If you grow up in, say, North America, you're likely to get used to
seeing the Sun rise once a day in the East, and be tempted to assume
that it does this all over the planet.  If you're into science or
science fiction, you might add a mental caveat that it is actually the
horizon moving, not old Sol.  Because our daily experience is so
uniform, it's hard to remember that at the North Pole the Sun rises in
the South, and does so only once per year... and that if most humans
happened to live there, "the Sun rises in the South, once a year"
would be the common wisdom.  If it seems ridiculous to you that any
human community could possibly be so parochial as to suppose that the
Sun behaves over the entire planet as it does at the North Pole, I
invite you to consider the concept "simultaneous".

In the universe we live in, in a general setting, there is no such
thing as "simultaneous", and Newtonian mechanics generally is absurd.
These ideas are useful in practice because humans currently live
in a very peculiar arrangement:
 * They think glacially slowly;
 * They all live in essentially the same place;
 * They are all going essentially the same direction at the same speed.
If you relax any of these conditions, "simultaneous" ceases to mean
much.

To draw an analogy, if Terran civilization were based on sentient
trees who all lived near the equator and faced East, we would not need
any concept of "North" and "South" as distinct from "left" and
"right", and it would be unambiguous to say "Tacoma is 35 miles right
of Seattle".  As it is, with people facing all directions at random,
"left" is in the eye of the beholder, and such a statement is fairly
useless.  Thus, we introduce the observer-independent notion of
"south", and say that "Tacoma is 35 miles *south* of Seattle".

In similar fashion, "simultaneous" is, for a general group of
observers, purely in the eye of the beholder.  Two events A and B may
be simultaneous to one observer, may happen in the sequence A,B for a
second observer, and may happen in the sequence B,A for a third
observer.  For useful communication to take place, an observer
independent vocabulary is needed.

Time, fundamentally, is about causality.  The most important
implications of "A happened before B" are that A may have caused B,
and that B could not possibly have caused A.

In the Newtonian approximation to reality, the speed of light is
infinite, interactions can happen instantly over arbitrarily large
distances, and a neat trichotomy holds: given any two events A
and B, either:
 1  A happened before B, and potentially influenced B.
 2  A happened after  B, and potentially was influenced by B.
 3  A was simultaneous with B.
Newtonian physics allows no other possibility.

Out in the real world, light needs a thousand picoseconds just to
crawl a foot, and things are more complex.  Given two events A and
B, there are fundamentally four observer-independent possibilities:
 1  Light from A could have reached, and potentially influenced, B.
 2  Light from B could have reached, and potentially influenced, A.
 3  A and B could have been co-incident -- same time and place.
 4  A and B could have been so distant in space and close in time
    that neither could have influenced the other.
In the Newtonian approximation, the latter two cases are indistinguishable,
but in our universe the distinction is vital.  Perhaps a couple of
examples?

Suppose we are debugging the Terrific Teraflops 2001 supercomputer.
The tracelog shows that a glitch at time/location B reversed the clock
polarity and fried 10E12 molecular switches, but why?  Another event
A, a nanosecond earlier and ten feet away, looks suspicious -- should
we investigate?  Newtonian intuition says "yes", since A was before B,
hence might have caused it.  Physics says "no", since A "4" B, and
could not possibly have influenced B.  To perform an automatic search
for candidate causes for glitch B, we would ask for all suspicious
events X such that X "1" B, not X "before" B.

Again, suppose you are wondering who murdered Ishtar.  If you are told
that one month prior to the murder, Thshpck broadcast an offer of
three megacredits to anyone who'd kill Ishtar, you might decide to
check out the local guns for hire.  But if you were also told that the
broadcast took place a light-year away from the murder, you could
immediately dismiss the broadcast as causally irrelevant to the
murder.

The bottom line is, any time you are interested in reasoning about the
universe we actually live in, the event-relationship categories 1-4
above are the ones useful for deriving valid deductions.

I don't really grok what follows, and anyone *really* interested in
understanding spacetime should pick up a good book on the subject, but
for those who want just a *bit* more and aren't allergic to a *gasp*
formula:

What's really going on is that the separate concepts of "distance" and
"elapsed time" need to be merged into the single concept of
"interval".  In the Newtonian formulation, an event A happens at point
Ax,Ay,Az, and independently at time At.  Similarly with event B at
Bx,By,Bz, Bt.  The distance between the two events is the Pythagorian

              2         2         2   0.5
    (  (Ax-Bx) + (Ay-By) + (Az-Bz)   )

and the elapsed time is independently

              2  0.5
    (  (At-Bt)  )

In Newtonian physics, it is assumed that the distance and elapsed time
can be computed independently, and that the results will be the same
for all observers.  No such luck!  If we insist on applying Newtonian
metrics to a relativistic world, we encounter the notorious
"relativistic contraction" of moving objects, &tc, and the
descriptions of events vary wildly from observer to observer.

To leave Newton for the company of Einstein, we must swap our
Euclidian space for a Minkowski space, and the above two independent
metrics for the single interval metric

              2         2         2         2  0.5
    (  (Ax-Bx) + (Ay-By) + (Az-Bz) - (At-Bt)  )


Even though independent observers will compute different distances and
elapsed times for a given pair of events, it turns out that (in flat
space and for freely falling observers, at least), the above interval
metric is observer independent, so multiple observers can again
sensibly compare notes.  Note that, while time now enters into our
"space" measurements, it does so with the sign reversed -- time is
still a distinguished dimension.

For two events which are distant in space and close in time, the
above square root will be of a positive quantity, the result will
be real, the interval will be "space-like", and causal interaction
between the two events is impossible.

For two events which are close in space and distant in time, the above
square root will be of a negative quantity, the result will be
imaginary, the interval will be "time-like", and causal interaction
between the two events is possible.  (So now you know what an
imaginary distance is :-)

Note that two events which happen at the same time and place are
separated by a zero interval, but the converse is not true: Two events
can be seperated considerably in time and space and still be separated
by a zero interval.  In the latter case, it is exactly possible to
send a photon from A to B, with no time left over.  What the two cases
have in common is that there is no time for thought between the two
events: any observer present at both events will perceive them as
simultaneous and co-incident... and in fact, the distinction between
the two cases is in the eye of the beholder, and won't appear in
a physically sensible tense system.

A quantitatively oriented language might relate the names for its four
tenses to the above arithmetic properties of the interval measure:

 past:           -imaginary
 future:         +imaginary
 here-and-now:   zero
 distant:        real

Simple tenses, in fact, might be simply permissible abbreviations of
a more complete quantitative interval measure.

(Loglan, of course, far from being a quantitatively oriented language,
follows in the anti-quantitative natural language tradition of
pretending that only black and white exist :-) :-) -- all functions
defined in the language must return TRUE/FALSE.  Were natural
languages designed by bigots to make others into bigots?  Only time
travel will tell :-) :-)

[2] by Bob LeChevalier (response to [1])

My first reading of Jeff P.'s tense commentary gives this (I may say more
upon further reflection):

1. Newtonian mechanics is a simplification of relativistic mechanics, and
in most natural linguistic situations works fine.  Therefore, by Lojban's
maximim optionality assumption, relativistic effects should be ignorable
if not important.
2. As Jeff notes, not everyone 'groks' relativistic tense, and it is 
irrelevant to most people.  It cannot therefore be a core concept that
MUST be learned in order to learn Lojban, unless either we come up with a
schema that makes it naturally and intuitively obvious.  Lojban's claim to be
a simple language depends upon this, and we can't realistically limit Lojban's
usefulness to those who understand relativity theory.  Nor can we be
expected to teach relativity to anyone learning the language.  Logic is
going to be bad enough.
3. Lojban tense structures are non-mandatory, and in fact are subordinate
claims that could be stated in a separate predicate.  Thus
   ko'a ba broda ko'e      x will razzlefratz y
is equivalent to  ko'a broda ko'e .i la'edi'u balvi ti
  x razzlefratz's y and this(the referent of the last sentence) is in the
future of this (i.e. here and now)

Tense structures must be simplistic, and can take only 1 sumti as incidental
information without horrendously complex grammar, or expanding into 
multiple sentences as in the example.  You can call this a grammar
limitation; it is usually acceptable.
4. If you expand the tense into a separate predicate, then relativity is
implicit in Lojban, and in fact is NOT limited to tense.  EVERY predicate
has the capability of adding, via sumti tcita, extra places, which can
include expression of causality and observer.  The latter two can also be
expressed separately in a 3rd and 4th predicate, if the sumti tcita
structure is too limited.
thus
ko'a broda ko'e .i la'edi'u balvi ti ga'a ko'i ri'anaiku
x razzlefratzes y, and this is in the future of here-and-now to observer z,
non-causally.
5. Note that Lojban separates causality from tense entirely.  THis is 
partly because Lojban has several kinds of causality, and you'd have to 
complicate Jeff's description by an indication of what kind did or did not
apply.  The example I gave used physical causality.

[3] by Arthur Protin (response to [2])

I would like to add my comments to this discussion, even though
I am not up to speed on learning lojban.

[4] by Jeff Prothero (response to [3])

Well, I'm not, either, of course... and have no ambition to be...

[5] by Arthur Protin (continuation of [3])

Many of the arguments on the relation of relativity and causality
are important and valid BUT probably should be dealt with by only
providing for clear and non-awful hooks.

[6] by Jeff Prothero (response to [5])

The specific proposal was just that Loglan should provide for all four
tenses, and adopt the physically significant definitions for "before"
and "after"... which coincide with the Newtonian definitions in the
Newtonian limit.  Should be at least as clear and non-awful as the
approach Loglan actually has, does, and will continue to take :-).
(As I stated in my original tense post, this is a personal pet peeve
which I bring up every few years.)

[7] by Arthur Protin (continuation of [5])

My understanding of our language
(which dates back to the 1970's) is that it is a language of appearances.
Thus the normal should be to assume a speaker frame of reference, with
provisions for adding things to shift that frame to:
    1) any arbitrary listener,
    2) a specified listener,
    3) an inertial reference,
    4) a time travelers reference,
    5) God's reference, and
    6) any other reference frame that can be described.

(that's not too much to ask, is it? :-)

I would expect than that it also makes sense to tie any notion of
causality to someone infering that relation, usually the speaker.

Thus, the simplest (least modified) lojban expression conveying
that "A happened before B" should mean that "as I witness/experienced
it A happened before B".  Some simple modification of that expression
should yield "As best as I can infer (from my knowledge of physics and
the events), A happened before B".  It should be possible with half
a dozen syllables to say "according to NAMEd paradigm, A happened before
B".  I expect that each of these types of qualification should be
available for all statements, or at least all observations.

[8] by Jeff Prothero (response to [7])

A most peculiar and unanticipated response, if I follow correctly --
you are arguing that Loglan should actually *favor* observer-dependent
descriptions?!  Thus, to take a specific case, you would *prefer* that
Loglanists say "Tacoma is to the right of Seattle" rather than "Tacoma
is south of Seattle" (with hooks so that it is easy to specify *which*
observer the observation is relative to), since this emphasizes 
"appearances", and makes the normal a "a speaker frame of reference".

Similarly, presumably, you would prefer that in Loglan altitudes be
given relative to the speaker rather than sea level, nearby sizes
should be given in multiples of the height of the speaker rather than
meters, mass should be expressed relative to the strength of the
speaker rather than in kilograms, distant sizes should be given in
terms of the percent of the speaker's visual field they cover, time
should be the local solar time (or time since breakfast?) rather than
speaker-independent timezone time, and so forth?

There's certainly nothing intellectually wrong with your proposal
(although, obviously, it runs diametrically opposite to the
Western-rationalist-scientific thrust of the last few centuries to
uncover and propagate objective observations which can be interpreted
and verified without reference to the observer!), but I'm curious as
to how you arrived at the conclusion that this is a part of the spirit
of Loglan... both the spirit of formal logic and JCB's aspirations to
"carve reality at the joints" seem to run directly counter to this
...?

(For what it's worth -- nothing, in my opinion -- JCB was very
enthusiastic about an earlier "relativistic tenses" piece I did for
The Loglanist, and it was only with difficulty that I got him to pull
it when Keith Wright (?) correctly pointed out that the equi-interval
curves I had drawn as circles should in fact have been hyperbolas...)

[9] by Bob LeChevalier (response to [8])

The default observer reference frame in a lojban sentence is unspecified,
and hence vague.  sumti tcita can be used to add places to any predicate
to specify an observer (ga'a) or a frame of reference (mai) or both.
I think this will suffice for all of Prothero's posed problems and Arthur's
specific examples.

[10] by Arthur Protin (response to [8])

(Your message makes you sound like such an extremist. ;-)

[11] by Jeff Prothero (response to [10])

Yes, this is perfectly true, I'm about three
sigmas off the norm.  E.g., I still like a priori languages, I think a
human language based on arithmetic would be interesting, I think
Loglan has no chance of beating Esperanto at its own game, and I think
"human" and "person" are orthogonal concepts... :-)

[12] by Arthur Protin (continuation of [10])

I must clarify something:  "Appearances in an observer reference frame"
IS NOT identical with "observer dependent".

[13] by Jeff Prothero (response to [12])

Maybe you must, but you don't :-) !  If an observation is relative to
an observer reference frame, then it cannot be interpreted without
knowledge of the observer... which is extra work and a hassle, hence
the universal tendency inside and outside the sciences to
communication real-world information in observer-independent terms.

[14] by Arthur Protin (response to [13])

Sure, but there are a great many ways that a report can be observer
dependent without reflecting what was apparent to the observer.

[15] by Arthur Protin (continuation of [12])

When I was in Seattle, I did not observe Tacoma to be to the right
(are they really that much more conservative there.)  I did find it
to be "before" Seattle.  But that had only to do with how I approached
it.  The only observer reference frame descriptor that accurately
describes to the observer that orientation necessary to observe Tacoma
while in Seattle is "South".

[16] by Jeff Prothero (response to [15])

Which statement is only useful given more information about the
observer, as you note.

"South" is a completely observer-independent concept; the
statement "Tacoma is south of Seattle" remains true no matter where
the observer is or what direction s/he/it is facing.  The relativistic
"interval" measure has the same property: it remains the same for
observers with different positions and velocities.  "Time" and
"distance" measurements don't have this property.  Thus, observers in
a relativistic context will prefer to report intervals to times and
distances for exactly the same reasons that you prefer to report
"South" to "right" in the case of Seattle/Tacoma.

[17] by Arthur Protin (continuation of [15])

As far as altitudes are concerned, I am currently at a loss to define
the "normal" because I am totally oriented toward the instrumentation
that measure altitude.  I never considered it before I was taught to
think of it relative to sea level.

Sizes should never be given in multiples of the height of the speaker
because, if my experiences are representative, the speaker can not truly
OBSERVE his/her own height.

[18] by Jeff Prothero (response to [17])

You don't have to "observe your own height" to notice that the rock is
twice as big as you, do you?  (It's only the poor auditor who must
scurry around doing extra work because of your laziness...)

[19] by Arthur Protin (response to [18])

You most definitely do.  You can only crudely approximate the comparison
of anything else to your own height, but you can do much better at comparing
the height of something with the height of something else placed near by.
My use of rulers (in either feet+inches or meters) is not dependent on
my desire to communicate those measurements to someone else.  I also use
measurement devices where their sole benefits are precision, accuracy, and
repeatability.

[20] by Arthur Protin (continuation of [17])

Trying to answer each instance of your misinterpretation, reminds me
that we are taught how to observe and how to express those observations.
None the less, the language design, as it was presented to me all those
years ago, was that it was not to embed the moral, religious, or
philosohical models but to be based on the appearances of the common
experience.

[21] by Jeff Prothero (response to [20])

Which is to say that it should be observer independent -- interpretable
without knowledge of the moral, religious or philosophical properties of
the observer.

[22] by Arthur Protin (continuation of [20])

As much as my beliefs are scientific, and I want to bias my childrens
thoughts toward scientific models, I am willing to give up biasing
the language to my beliefs in order to keep it from being biased by
anyone elses beliefs.

Thus a chair is a chair not because it has the soul of a chair but
because it appears that it can function as a chair. Recognizing it
as a chair is the act of the observer.  I have had the experience
of not recognizing a chair as a chair because it accomplished its
function in a manner that I had not seen prior.  It became a chair
(something I could sit on) only when the observer (myself) could
visualize how it could function as a chair.  My friend's chair is
not necessarily a chair to me.

[23] by Jeff Prothero (response to [22])

I think you interpret this opposite to JCB!  I think JCB was trying to
come up with an observer-independent description of chairness
dependent only on the publicly-available properties of the chair
itself, whereas you seem to be headed for a private notion of the
predicate which is uninterpretable execept with respect, not just to a
particular speaker frame of reference, but in fact to a particular
speaker.  Now Loglan is descending into a Babel of private languages!

[24] by Arthur Protin (response to [23])

This line of reasoning sure seems like a mess now.

A "description ... dependent only on the publicly-available ...." looks
like a good thing but it has the key to the problem.  Where is this
definitive repository of "publicly-available" information?  And is the
public really going to avail itself of that availability?

[25] by Jeff Prothero (response to [24])

Physical phenomena, as distinct from cultural convention.  E.g., "sphere" and
"cube" can be distinguished in culture free fashion, but "freedom
fighter" vs "terrorist" (or "authoritarian" vs "totalitarian") can be
distinguished only by phoning the Ministry of Truth.  (To pick a
topical example, Saddam Hussein moved from being a genial
authoritarian leader to being a dastardly totalitarian dictator when
he switched from invading Iran to invading Kuwait. Alien observers
equipped only with satellite photography might be hard-pressed to
distinguish the two actions...)

I think JCB wanted to keep the "sphere"/"cube" distinctions in Loglan,
but stamp out the purely connotative distinctions.  This would make
much American political discourse virtually untranslatable into
Loglan, of course, since "Joe believes the Evening Star is the Morning
Star" just isn't the same as "Joe believes Venus is Venus" :-)

[26] by Arthur Protin (continuation of [22])

Rather than try to perfect the model of the universe before defining
how to talk about it, I back the approach of defining the language to
be as free of the model as possible and to make the role of the model
explicit such that evolving and conflicting models can coexist for
a sufficient length of time for the better one to prevail.

[27] by Jeff Prothero (response to [26])

I think this makes your point of view clearer... language seems to me
inherently unsuited to this role, cinema verite' seems much closer to
presenting a raw, uninterpreted sense-stream for the audience to
interpret as they wish... communication of data with no transfer of
understanding. (Personally, I think transfer of understanding is
fascinating and laudable :-)

[28] by Arthur Protin (response to [27])

I like trying to transfer understanding as well, but there appears not
to be any way of assuring that that transfer occurs.  Communication seems
to be a probabilistic endeavor which is very dependent on the speaker and
listener sharing a great many common experiences, enough to have developed
shared common abstractions and models.

I don't have a problem with providing for the additional tenses needed
for relativity, but will the expansion exhaust the pool of available
tenses.  How will we provide for the many strange ones that are required
for discussing time travel?

[29] by Jeff Prothero (response to [28])

(Only one additional tense, in my formulation.)  I seriously doubt that
time travel will become a real-world problem in a physical sense, but
if Loglan were to replace Latin as the standard funny-looking SF language,
it might conceivably become a real-world linguistic problem :-).  Comments:

* Since they are pure fantasy, there are as nearly as many models of time
  travel as there are writers about it.  Hard to cover them all.

* I once proposed, and RLC agreed (a unique occurrence ? :-) that Loglan
  should reserve a "do not define" bank of LittleWords for experimentation.
  These can be used for local-to-a-story tenses.  E.g., you might have
  separate tenses for separate parallel worlds, or for self-created
  causal loops ("All You Zombies...").  (To the folks who will immediately
  note that English time-travel SF gets by fine without extra tenses:
  Maybe the stories that really need them don't get written.)

* The difference between Newtonian space and Newtonian time is precisely
  that you can travel freely through one but not the other.  In the limit
  of Newtonian timetravel, time simply becomes a fourth spacial dimension.
  This suggests that in the limit, you should simply drop the tenses and
  use the spatial-position LWs.

[30] by Bob LeChevalier (response to [29])

We already allow for 4-dimensions in the spatial tenses.  Does this mean
that we've covered your relativistic problems?

[31] by Jeff Prothero (response to [30])

4 dimensions? How did you justify LWs for that?  Not really, the
limit is a long way away.  In the limit, a circle becomes either a
line or a point, but it's still nice to have a word for circle :-) But
it does count as a point against time travel. 

[32] by Bob LeChevalier (continuation of [30])

Can anyone come up with a REAL example of a sentence that would use this
hypothetical 4th tense, including sufficient context as who is saying it
to whom, and why none of the other tenses are correct, provided that
the tags 'to observer ...' and 'in reference frame ...' (ga'a and ma'i)
are specified.  Can you argue that the examples of such usage are numerous
enough to justify use of a cmavo or in a very limited
list of available words left?  Can this '4th tense' be predicated instead
of abbreviated with a cmavo (all tenses are considered to be abbreviations
of some subordinate predication, as jimc is oft wont to point out)?

[33] by Jeff Prothero (response to [32])

I'd hoped my computer example provided such a context... some further
comments:

It's always *possible* to do it with epicycles, the Copernican model
only wins on elegance.

The argument for 4th tense as LW comes more from symmetry and elegance
than Zipf. If you are heavily using a set (simple tenses) with exactly four
members, it is inelegant (and potentially misleading) to have names
for only three of them.

Fortunately or otherwise, arguments from elegance are more likely to
move mathematicians than businessmen.

 It is definitely unfortunate -- IMHO, a design blunder -- that Loglan
is designed around small, fixed-size lexeme slots that easily
overflow.  It would not have been difficult to make all lexeme slots
infinite, so that the number of defined members of each lexeme could
be governed more by logical and physical necessity, and less by the
structure of the human vocal tract. A good exercise for the Handbook
of Human Language Engineering.

[34] by Bob LeChevalier (response to [33])

The fixed size is based on a limited word space for the cmavo.  The predicate
word and borrowing word spaces are infinite.  It is arguable that far more
than half the cmavo we have now, including ALL of the tenses, are Zipfean
abbreviations for predicates.  A limit on abbreviations seems much less
offensive than Jeff's accusation, now, doesn;t it?

Yet we are fully prepared to add to certain lexemes, and some lexemes have 
equivalents that may be concocted ad hoc from the predicat words (including
tenses.  So if you can express the relativistic and/or time travel concepts
as predicates, we can always use this method to cover them for tenses).  Thus
even some of our cmavo lexemes are effectively open-ended.

This sounds like an argument for making the language some kind of perfectionist
ideal TOTALLY DIVESTED FROM THE BASIC PURPOSE OF LANGUAGE, which is human
communication.  If human vocal tracts (and hearing and multi-track processing
and other limits caused by our being human) are not allowed for in the language
design, it is NOT a human language.  Or do I misunderstand your argument, Jeff?

[35] by Jeff Prothero (response to [34])

(I *don't* propose changing Loglan's design in this respect.  I *do*
propose that follow-on designs try to do better.)  You say the glass
is half full: I say it is half empty :-): something approaching half
the cmavo aren't such abbreviations!  The point is just that, if
you're starting from scratch, you can make all word classes
variable-length without much effort.

I think you do!  My point is that (e.g.) the number of logical
connectives in your language should not be dictated by (say) the
number of vowels humans can clearly enunciate.  It is perfectly
possible to design a language with such a constraint (and 1960 Loglan
had a flavor something like this, no?), but it is also perfectly
possible to design a language *without* such a constraint, and
have it be just as human-usable.

Riding the point to one of my favorite outre' ideas: It is perfectly
possible to design a pronounciation system for your alphabet which
makes all sequences of letters equally pronouncable. This approach
lets word design take a much more logical course, without irritating
special cases due to the fact that your regular word-construction
algorithm produced an unpronouncable result.  My own experimentation
goes further and adopts an arithmetic basis... but we're wandering
away from Loglan.

My basic point was just that Loglan's word morphology means that we
are forever in danger of painting ourselves into a corner somewhere in
LittleWord space.  It is always possible to get out of such a corner,
at the cost of a few footprints, but a better initial plan could have
eliminated the problem once and for all, no?  Language prototypers
take note.  (This belongs on a Language-Design mailing list, but
there's not enough traffic to justify one yet, and lojban-list seems
the closest current approximation.)

[36] by Bob LeChevalier (continuation of [32])

In short, I haven't seen why this is needed, and the discussion doesn't seem
to provide enough information on how it would be used in order for any
teacher to teach it.

[37] by Jeff Prothero (response to [36])

Well, I think Jim considers this topic just short of too trivial to
discuss, Art thinks I'm probably but not yet provably daft, and you
think this is a waste of valuable time.  Maybe if I can convince Art a
fourth tense is worthwhile, we can hammer out something specific
enough to satisfy you... :-)

[38] by Bob LeChevalier (continuation of [36])

There is a different standard held up for assignment of cmavo (LWs) than
for brivla (Predicate words) in Lojban.  We can easily devise words as
brivla, even on the fly.  People who don;t undersatnd can ask the speaker
what the word means.

cmavo, though, have to have a grammar assigned to them (lexeme or grameme),
possibly have a rafsi (affix) assigned, and take up some of a very limited
word space.  In addition, they appear in a dictionary, as well as in LogFlash
and must have a definition.  (The same is true, of the gismu - the root
'primtives', by the way.)  These lists will be baselined, and we are planning
in even the earliest books to have examples of the use of each word, so that
people have enough sense of what they mean to use them, since we expect them to
memorize them.  They will prefereably BE useful, and have a definable use.
Theoretical usages that cannot be explained are garbage.

(This is not to say that all possible Loglan expressions have to have a
pre-defined semantics.  It means that we have to present enough information
in our prescription of the language that someone who wants to figure out the
semantics for themselves, can do so.)

[39] by Jeff Prothero (continuation of [27])

(I think) there's an inherent quantitative problem in your attempt to
squeeze the model out of language: human perception supplies hundreds
of millions of bits of data per second, but human language conveys only
about ten bits of data per second, so any human reporter is necessarily
providing only a very abstract (== model-dependent) transcription.

By breaking a continuous world up into discrete objects, and then
further breaking this continuous spectrum up objects into a few
discrete classes of objects (predicates), a human language is (IMHO)
necessarily imposing a very strong model on the universe -- a strong
base for the Whorf Hypothesis.  In the case of tenses, which are
particularly deeply embedded in the language, I thought it might be
nice to have the model be a physically sensible one instead of the
traditional fiction :-)

By "the better model" you mean whichever one in fact prevails, right?
Or do you have a more fundamental measure?)  By this logic, it seems you
should favor a fourth tense, since it lets the Newtonian and Special
Relativistic tenses compete on an even footing.  Or would this require seven
basic tenses so that the auditor always explicitly knows which model the
speaker is using?

[40] by Jim Carter (response to [27])

Physically you have no choice but to use observer-dependent
descriptions, since the act of observation randomizes what is observed.
Of course in everyday life there is so much correlation, in time and
space, between the randomized part of a system and the remaining parts
that you can get away with a statement "The creature is moving toward
you" as opposed to "I saw the creature's left upper tentacle move
toward you just now".

[41] by Jeff Prothero (response to [40])

(Um, we're wandering from relativity to quantum...)  I'm not sure this
has anything to do with observer-dependence.  The fashion and degree
to which the observed system are randomized depend on the system and
the probe, not on any characteristic of the observer who launched the
probe.  It means you have are guaranteed a certain minimum amount of
problem reconciling successive observations of a system, but this
isn't observer-dependent: you have exactly the same problems whether
the successive observations are collected by different folks or the
same observer.

[42] by Jim Carter (continuation of [40])  

In talking about emotion, we're all aware that people perceive the same
event differently and so many (all?) of the Lojban root words about
emotions include an explicit case for the experiencer; you don't even
have to use a modal phrase ("sumti tcita", yucky terminology).

As for relativity, the reference frame is in fact irrelevant for a
causal connective.  Starting at a "causative event" X, the universe is
divided in three parts by the null (lightwave) cone centered on X into
X's future (events reachable at less than or equal to lightspeed, which
it could conceivably cause), X's past (which could reach it at less
than or equal to lightspeed), and X's "present", which can neither
cause nor be caused by X since the required communication would be
faster than light.  The null cone is invariant under arbitrary
coordinate transformations anywhere (because the metric is a tensor)
and hence choice of a reference frame cannot move an event between
these three categories.  It's true that any "present" event could have
a time coordinate before or after X depending on the frame, but that
sign is irrelevant causally; it's still a "present" event.

[43] by Jeff Prothero (response to [42])

Your interpretation is defensible, but not the only one and (to my
eye) not the most natural.  In English, at least, when one says that
a person (or thingie) is "present", we mean precisely that it is
practical to interact with that person (or thingie).  Your definition
reverses this, and defines the "present" to be that portion of spacetime
with which we *cannot* interact!  

Perhaps my viewpoint becomes clearer if we reflect that the "causative
event" X is in practice usually not a spacetime point, but rather a
volume.  One can justify this by noting that most real-world events,
such as a bullet shattering a window, or the fall of the Roman Empire,
have a nonzero duration in time and extent in space.  (Could one also
make this point by an appeal to the Uncertainty Principle?)  This
makes the null cone less of a negligible boundary (as you are treating
it) and more of an interesting volume in its own right -- the
"present", in my four-part treatment.

In this treatment, the "past" and "future" are as you described, but
the "present" is the null cone, and (for non-pointlike events X)
includes that set of spacetime points which the event (potentially)
*interacts* with: both (potentially) influences, and (potentially) is
influenced by.  (Your "present" becomes my "fourth tense".)  To
modify your definition:

Starting at a "causative event" X, the universe is divided in four
parts by the null (lightwave) cone(s) centered on X:

 X's "future":  Those events reachable at less than lightspeed,
                which it could conceivably cause;
 X's "past":    Those events from which X could be reached at less
                than lightspeed;
 X's "present": Those events reachable from X, or from which X can
                be reached, following a geodesic at lightspeed,
                -- the null cone itself.  For a non-pointlike event X,
                this fuzzes from a (hyper) surface into a (hyper) volume
                containing events which X both potentially influences,
                and also by which X is potentially influenced.
                (One can also look at this as the set of points
                composing X, or the set of points at which X is present,
                or the set of spacetime points co-incident with X for
                some possible observer.)
 X's "????":    Those events which which can neither cause nor
                be caused by X since the required communication
                would be faster than light.

(Should we call "????" X's "heaven"? X's "ignorance"? X's "protopast"?
X's "exfuture"? ... :-)

[44] by Jim Carter (continuation of [42])  

It would be satisfactory to me to define Lojban "present" to mean
"events separated from the reference event by a spacelike interval". 
In daily life the thickness of the present is too small to observe,
so Lojban would not be spoiled for daily life, yet a precise definition
is there for applications, like ultra-high-speed computers and space
travel, where speed-of-light delay becomes significant.

[45] by Bob LeChevalier (response to [44])

Per Prothero's comment on English meaning of 'present', I think that the word
has split into 2 (or more) meanings, and he seems to be jumping between
them.  When, in English, "someone is present", this means indeed that
you can interact with them, but this is due to colocation in SPACE (and
only incidentally in time).  The synonym is 'here'; i.e. "someone is here".

[46] by Jeff Prothero (response to [45])

Well, as I see it, at least, the question is how to extend the English
"present" tense from the Newtonian limit into a more general setting.
Since humans, to date at least, haven't had to deal with relativistic
effects a lot in daily life, this is necessarily as much of a creative
act as a logical one...

The meaning of a word can be defined denotationally as a set of
situations (as Montague mostly does), but is most interestingly
described as the set of relevant reasoning rules: knowing that
a certain word describes X, what are we entitled to conclude?

We (I?) would like to generalize the current definition of the present
tense so as to preserve the sorts of reasoning now possible with it.
(Since relativity views the world in terms of events and lightspeed
communication, it would be nice to frame our definition in that
language.)  What is the most essential difference between "exists",
"existed" and "will exist"?  I was arguing that it is the potential
to interact:

If someone tells me that the Mississippi *is* a great river, I feel
entitled to suppose that I can, at least potentially, go visit it.

If someone tells me that Gibraltar Falls *was* one of the greatest
scenic wonders on the planet, I'm rather inclined to conclude that
it's too late to see them. (Last I heard, it's currently believed that
there never was a Gibraltar Falls.  Sigh.)

If someone tells me that WW III *will be* impressive to watch from
Luna, I'm inclined to believe that nobody has seen it yet, and that
I can't count on just hopping a shuttle to enjoy the show.

On this reading, the English present tense includes events out to
infinity in all directions precisely because, in the standard
Newtonian model, you can potentially interact with anything "currently"
existing, if only you get there fast enough.  (This intuition seems so
strong that people appear to somehow feel robbed by the lightspeed
barrier, that it must be a cruel trick, that somehow getting out to
everything out there "now" must just require a little engineering.
Conservation of energy doesn't seem to arouse quite the same
indignation. Is this because English tenses don't suggest you *should*
be able to get something for nothing? :-)

Following this reading, the essential semantic property of the present
tense which should be preserved in its new incarnation is that it
selects those events which we can potentially interact with.  Which is
more or less what my definitions have attempted, and which JimC's
latest paraphrase may have done most neatly.

The "problem" with this definition is that it leaves a (vaguely
donut-shaped) hole in spacetime, consisting of those events which are
neither past, present, nor future, for which I need a fourth tense...

[47] by Bob LeChevalier (response to [45])

Now I can recognize a whole bunch of arguable connotations about time that
can certainly support "present" in the sense of here matching "present"
in the sense of "present tense", but I for one don't think of them when
I use the term in the 'here' context.  In Lojban, we have the two gismu
	'zvati' x is present/'here' at    
  and
	'cabna' x is simultaneous with 
for the two meanings.

(I recognize that I may have totally misunderstood both Prothero and Carter in
your usages of 'present', but I never made it through relativistic physics,
which is why my degree in astrophysics has made me 'only' a computer-type.)

In any event, we have to teach the language, and relativistic effects are
a bit too tough for any textbook I'll ever write.

[48] by Arthur Protin (response to [45])

An example that might illuminate this noise (warning: pun in progress)
is the phenomena of lightning and thunder.  Two separate manifestations
of one event, which due to their propogation delays reach us at different
times (actually due to the assignment of terms before understanding the
event, lightning refers both to the flash of light and the discharge that
causes the flash with a bang).  Is thunder the sound wave that starts
simultaniously with the flash, or is it the experience of hearing that
sound many seconds after the seeing the flash.  I consider it to be the
latter (the hearing) extrapolated to the unobserved formation of the
sound (the observation is of the wave radially propagated many seconds
(and thankfully dissipated by that radial propagation)).

Do the flash and the bang occur at the same time?  No, and we would not
describe them are simultanious, but we all
understand that they originate together.

[49] by Jeff Prothero (response to [45])

I think you're panicking before the attack is sounded, Bob!  Loglan
can have a word for "twistor" without requiring you to teach quantum
mechanics in the textbook, it can have a word for "gravitational
lens" without requiring you to teach General Relativity in the
textbook, and it can have a fourth tense without requiring you to
teach Special Relativity theory in the textbook.  Heavens, we can even
have a word for "pulsar" without covering stellar evolution!  As long
as you keep your textbook examples in the realm of Newtonian
mechanics, the current three primitive tenses remain unchanged.

As I see it, at least, the subjects of the tense discussion are

(1) Should Loglan define the tenses *away* from the Newtonian limit? and
(2) Does the general Special-Relativistic case require additional
    tenses?

I think at least two or three other folks believe (1);
I'm not sure if I've yet convinced anyone else of (2).

Since I've been through this discussion a couple of times before
without having the slightest influence on Loglan, I'm not really
holding my breath in expectation of a fourth tense.  Might pop up in
Guaspi, though, since Jim understands this stuff considerably better
than I do...


[50] by Jim Carter (response to [49])

The four-part division is interesting for extended events,

An Earth-bound person would be tempted to define "present" as those points
in the future relative to parts of the event and in the past relative to
other parts.

[51] by Jeff Prothero (response to [50])

Sounds like a concise, precise definition.  (Maybe that's why you're in
math and I'm in anatomy :-)

Drek. Maybe we need to go to *five* tenses.  In my previous post(s) I
was very careful to say only that my fuzzed-out null-cone "present"
*included* those events which could interact with (both influence and
be influenced by) our extended event X, since this "present" clearly
doesn't *consist* of such events -- it includes many more.  I'm
beginning to think we need to distinguish these two "present"s:

Coincident-present
------------------

Given an extended spacetime event X, the "coincident-present of X"
consists of all (non-extended) spacetime events Y such that (some part
of) X is co-incident with Y for (some possible) observer.  This is the
union of the null-cones emanating from X.

Interactive-present
-------------------

Given an extended spacetime event X, the "interactive-present of X"
consists of all (non-extended) spacetime events Y such that
Y is both in the future of (some part of) X and
          in the past   of (some part of) X
This is the set of events which can *interact with* (both influence
and be influenced by) X.  This is a proper subset of the
coincident-present, and corresponds the most closely with the normal
English/Newtonian "present" (IMHO).

[52] by Jim Carter (continuation of [50])

But when you start charging around at a substantial fraction
of the speed of light, the volume which you have to consider which is
neither past nor future becomes the majority.  Is he coming in range
and I don't know it?

[53] by Jeff Prothero (response to [52])

I'm afraid I'm not following you.  Presumably we can treat ourselves
as being at rest, so your "when you start charging around at the speed
of light" can be treated equivalently as "when you start interacting
with objects travelling near lightspeed".  (Our "Earthbound person" is
encountering photons, of course, but not commonly interacting with
them as enduring and individual objects.)

But I don't see how the volume which is neither past nor future can
"become the majority" (of spacetime, by "volume", I presume) unless
the extended event itself occupies the majority of space or time.
(Taking myself as an example: An extended event (should we make
"thing" a technical term? :-) a few feet (light-nanoseconds) by a few
decades (hundreds of megaseconds) across will have an
"interactive-present" a few light/decades in radius no matter how you
cut it...  in a universe a light/giga-decade or so in radius.  Since
any chord through this "present" is an observer-invariant interval, I
(naively?) presume the "volume" of this "present" is also
observer-invariant, and always much smaller than spacetime.

The "coincident-present" of the above extended event will be much
larger (one can visualize it as a decades-thick sphere expanding
outward through spacetime into the past and future, no?) but still
fills only an infinitesimal fraction of spacetime.

Perhaps you are thinking of the "present" as including all of any
extended objects which pass through it?  I'm not -- my "present" is
*defined* by an extended object, but *consists* of classical
point-events.  The elder George Burns is in my "present", but not the
young one.

[54] by Jim Carter (continuation of [52])

Similarly in real life, a conversation would consist of a timelike, not
spacelike chain of sub-events (speeches), and it's neither useful nor
polite for both parties to transmit at the same time.  Relative to the
last speech, the union of your two present zones are points where the
listener should be listening, not talking.

[55] by Jeff Prothero (response to [54])

(Well, there was a time when they sang four part harmony with
different words, even in different languages, for each part! :-)

My apologies: I'm afraid your point completely escapes me.  What do
the pragmatics of conversation have to do with tenses, relativistic or
otherwise?  Where did *two* present zones come from, here?

[56] by Arthur Protin (response to [43])

Now that I am really starting to think about tenses as verses thinking
about relativity, I find I am even more confused by your proposal.  Please
give me examples of the tenses that you propose opposite the tenses that
exist right now complete with the transliterate English.  These examples
should be complete phrases if not sentences.

[57] by Jeff Prothero (response to [56])

Well, my bits on the computer glitch and space murder were supposed to
have provided such examples.  I'd have coded up some specific Loglan
phrases, but I don't even know what "pa na fa" have become these
days...

Let's return to the computer example.  We are talking to each other at
time T, trying to deduce why the prototype BioFlow 2000 computer just
fried it's crossover, by searching the various RNA logs.

The immediate cause of the failure is obvious -- a glitch G reversed
the clock phases, resulting in a dead short in all active crossover
switches.  We're both aware that such a glitch is "utterly
impossible", of course, and are trying to figure out what might have
led to it...

You discover a fluke F on a related RNA log, and we discuss it
relative to glitch G:

[1] "Look, this module had just fluked!"

[2] "No, it was just *going* to fluke -- check the logtime!"

[3] "You're forgetting, this log had been running ten nanoseconds slow
for nearly a millisecond -- the module had been fluked for two
nanoseconds at this point."

[4] "Oh, right.  But look -- the module is ten feet
[light-nanoseconds] from the crossover.  The module *tense-four*
fluked.  Any other flukes?"

(These are all compound tenses in English since in speech we're
required to do a bank shot from time-of-speech off time-of-glitch to
get to time-of-fluke.  In at least some loglans, these could be
simple tenses because we would just adopt G as our time of reference.)

Statement [1] notes that F was before G, using the past tense.
The unstated implication is that F might have caused G, hence
deserves further investigation.

Statement [2] protests that F was after G, using the future tense.
The unstated implication is that F couldn't possibly have caused
G, and might in fact have been a consequence of the general
collapse of the machine after G, hence isn't worth wasting any
time on.

Statement [4] points out that, even though F was before G in a
wall-clock / Newtonian sense, F was further away in feet than it was
previous in nanoseconds (light travels almost exactly one foot per
nanosecond, as Grace Hopper observes), hence F lay in the
"oblivious zone" of G, and could neither have influenced, nor
have been influenced by, G.  Note that neither past, present, nor
future tense could make this point properly!

[58] by Arthur Protin (response to [57])

Thank you for the example because it starts to provide evidence that
the "special" tense that you ask for may be unneeded.

I finally saw what you were asking for a distinction between present tense
interactable and present tense non-interactable.  Thus, the all-important
fourth line of your example is not correctly represented in English.
It should be a compound of the past tense with the present non-interacting,
more like: "The module was independently fluking".

[59] by Jeff Prothero (response to [58]) 

Right, I should have said "The module had *tense-four* fluked".

[60] by Guy Steele (response to [59])

No, I think it should be "The module did *tense-four* fluke."

[61] by Jeff Prothero (response to [60])

Maybe! I examined:

 "had fluked"               (                 at past reference point)
 "had been going to fluke"  (future relative  to past reference point)
 "had been fluking"         (continuing       at past reference point)
 "had been fluked"          (past relative    to past reference point)

[62] by Guy Steele (response to [61])

No, this last one is a case of passive voice.  It is confusing
that the participle and the past tense have the same form.
Your table should read:
    "has fluked"               (                 at past reference point)
         or "fluked", or "did fluke"
    "had been going to fluke"  (future relative  to past reference point)
    "had been fluking"         (continuing       at past reference point)
    "had fluked"               (past relative    to past reference point)
Right?

[63] by Jeff Prothero (response to [62])

I grab my American Heritage Dict to look up "participle", "gerund",
"nominative" ... then carefully re-read:]

Right!  I stand corrected.

[64] by Jeff Prothero (continuation of [61])

The pattern appears to be: "add a 'had' to reach past reference point",
with various operators added to refer to the past, present or future
of that reference point.  So I extended the pattern with:

 "had *tense four* fluked"  (oblivious zone relative to past refpoint)

with the trailing "-d" probably inappropriate, but an aid to intuition.

[65] by Guy Steele (response to [64])

The problem is that English uses both suffixes and pre-positional auxiliaries.
I chose the "did" form so that everything could be consistently cast as
auxiliary operators in the same left-to-right order that Lojban uses.

[66] by Jeff Prothero (continuation of [64])

It's not clear to me that one can do the same sort of temporal
bank-shots quite as naturally with "did":

 "did fluke"                                    (present-of-past)
 "did already fluke"                            (past-of-past)
 "did ready to fluke" ("did about to fluke"??)  (future-of-past)
 "did *tense four* fluke"                       (obliv.-zone-of-past)

(Obviously *all* of these examples misuse the noun "fluke" as a verb.
This is not an unrealistic extrapolation, given the steady
Haigspeakification of the language... and it simplified the example
dialogue.)

[67] by Guy Steele (continuation of [60])

Hm.  My next thought was that if one can compound a past tense with
tense-four, then such auxiliary verbs as "had" also ought to come in
four flavors (that would fall out naturally from the orthogonal
structure of Lojban, I hope, but it bears pointing out explicitly for
the English-pidgin version).  So I ought to be able to say, e.g.,
"The module *tense-four* *tense-four* fluke"; in other words, the
module, in an event tense-four-related to the speaker, has a fluke
event tense-four-related to the event under discussion.

[68] by Jeff Prothero (response to [67])

Yes.  In Old Loglan, we had

 pa (past      zone)
 na (present   zone)
 fa (future    zone)

to which I can add for purposes of discussion

 qa (oblivious zone)

yielding the Old Loglan psuedo-examples (I just invented the predicates)

 "le modlu na     fluke"  The module is fluked/fluking.
 "le modlu nana   fluke"  The module is now fluked/fluking.   (?)
 "le modlu pa     fluke"  The module was fluked/fluking.
 "le modlu pana   fluke"  The module was then fluked/fluking. (?)
 "le modlu pafa   fluke"  The module was about to fluke.
 "le modlu papa   fluke"  The module had already fluked.
 "le modlu papafa fluke"  The module had already been about to fluke.
 "le modlu qa     fluke"  The module *tense-four* fluked/fluking.
 "le modlu paqa   fluke"  The module had *tense-four* fluked.
 "le modlu paqaqa fluke"  The module had *tense-four* *tense-four* fluked.

Old Loglan syntax allows the simple tenses to be chained as deep as
the speaker desires.  I don't know the details of the current Lojban
treatment, but I'm confident it is similar in spirit. (Bob?)

[69] by Guy Steele (continuation of [67])

But my third thought was that such sentences, though conceptually useful,
ought to occur only in fantasy and science fiction, because the speaker is
making a claim about an event with which he cannot have interacted
and therefore of which he can have no knowledge.

[70] by Jeff Prothero (response to [69])

Certainly, you cannot *know* what is currently happening in your
oblivious zone, that being a defining property of the oblivious zone.

[71] by Guy Steele (response to [70])

Actually, that was the very core of my observation: that you cannot
influence it at all.

[72] by Jeff Prothero (response to [71])

(Ok, so maybe you want to banish the past tense to SF? :-) :-) 

I think we can agree that of my four spacetime zones (past, present,
future and oblivious), the oblivious zone is of the *least* immediate
importance -- we get information from the past, interact with the
present, and try to influence the future, but interact with the
oblivious zone only one step removed, by guessing influences our past
actions may turn out to have exerted on it, and anticipating
information we may eventually receive from it.

[73] by Jeff Prothero (continuation of [71])

But you can extrapolate such activities, much as when you say "Aunt
Sue should be landing in L.A. just about now" (or as in my example of
a sister taking finals in a different solar system), and you can you
can talk in hindsight about what *was* going on in
someone's/something's oblivious zone, as in the computer-debugging
example under discussion.

If you think about it, I think you will find that your third-thought
objection applies with equal force to the future-zone tense and the
oblivious-zone tense, given that we are equally unable to have any
direct knowledge of events in the future or in the oblivious zone.
(The difference between the two, that we can hope to influence events
in the future, but not in the oblivious zone, doesn't appear to enter
into your objection.)  Do you seriously propose that the future tense,
although conceptually useful, ought to occur only in fantasy and
science fiction?

[74] by Guy Steele (response to [73])

I was overstating my case.  Maybe I mean that such a tense in English
would almost always necessarily appear in the subjunctive mood.

[75] by Jeff Prothero (response to [74])

A fixpoint: I agree -- the fourth tense will likely prove gregarious,
found rarely alone, usually with a subjunctive or in a compound tense.

[76] by Jeff Prothero (continuation of [73])

(This seems unlikely to me, leading me to suspect that I have yet
another person on my hands who is really arguing against the
unfamiliar.

While I'm making irrelevant comments: I will consider it
a real milestone when the fantasy crowd starts using relativistic
tenses!  Sort of like seeing the first personal-computer retail shops
appear, and hearing the first personal-computer radio commercials -- a
sign that the revolution has reached the mainstream.)

In general, I can easily imagine that keeping track of projected
events in your oblivious zone might be essential any time that you
*have* an oblivious zone... whether you are clearing buffer space (in
an amplifying, soliton-mode optical-fiber loop, perhaps? -- the spirit
of the mercury delay line may yet live!) for projected incoming
traffic from the optical net, or anticipating action from opposing
orbital laser battlestations (WW III may be essentially decided before
first news of the start *reaches* Earth :-( ).

[77] by Arthur Protin (continuation of [58])

As of right now I would not plead ignorance and abstain but would
vote a definite NO.  I have several reasons:

This change caters to the personalities prone to the logic flaw
"Post Hoc Ergo Poster Hoc" (It follows this, therefore it is caused by this).

[78] by Jeff Prothero (response to [77])

I think I understand your point, although I don't agree with it.  At
the risk of putting words in your mouth, you see me as unnaturally
slicing up the "present" to cater to some obscure technical needs
which, while perhaps valid, are uncommon, and which screw up everyday
reasoning in the process: The game isn't pointless, but it isn't worth
the candle.

In my opinion, of course, you are a Whorfian-mindlock victim who
prefers a familiar fiction to an unfamiliar reality. :-) If we peer at
the relativistic world through conventional Newtonian glasses, my
proposal does more than slice the conventional present into a near
(interactable) and far (non-interactable) present: these "presents"
(which I call the present and the oblivious zone respectively) widen
out with distance, eating into the naive Newtonian past and future.
(In the limit of infinite lightspeed (and zero reference-event
duration), this widening is negligible and the present collapses to an
infinitesimally thick cut through spacetime.)

The cost is a paradigm shift -- some non-monotonic learning -- and the
benefit is that statements made using the tenses carry information
more likely to be useful in everyday reasoning: IN A RELATIVISTIC
CONTEXT, the relativistic tenses allow the auditor to immediately make
deductions about potential causation -- deductions not possible if the
naive Newtonian tenses are used.  (In a Newtonian context, three
relativistic tenses become identical with the standard three, and the
fourth is unused... barring metaphorical use, which I consider likely
among folks who have made the mental paradigm shift...)

[79] by Arthur Protin (continuation of [77])

Interactions can be through a wide varienty of agents each with their
own separate properties, not the least of which is propagation rate.

[80] by Jeff Prothero (response to [79])

Yes, and while I haven't commented on this for fear of muddying waters
which most folks already find sufficiently opaque, it would be
possible to use relativistic tenses (and related predicates, of
course) in a loose metaphorical way when other communication rates
dominate the discussion.

In a world in which the fastest communication is by mail carried by
sailing ship, Africa is in London's oblivious zone for purposes of
day-to-day business.  As remarked above, I suspect a community of
people acculturated to the new tense paradigm would indeed find such
usages useful.

The difference, of course, is that such barriers are technological,
but the lightspeed barrier is not.  (The speed of sailing ships does
not enter into the ratio of mass to energy in a particle, for example,
but the speed of light does...)

[81] by Arthur Protin (continuation of [79])

While I accept relativity as the best expanation currently available
to us and its models are sufficient for describing those of our
experiences recorded (and reproducable) to date, I do not consider
it to be the last word on the subject. (I have already heard
theories that may be candidates to succeed relativity). 

[82] by Jeff Prothero (response to [81])

Are you still waiting to find out if the Earth is approximately
spherical? :-) :-) Modern science has an *immense* amount of
experience with the realm described by Special Relativity.  General
Relativity is fairly well tested, but this is very hard.  By contrast,
every particle accellerator, cosmic ray detector &tc built *must* take
Special Relativistic effects into account in order to work at all.  

If all of these gadgets turn out *not* to have been working for the
last century, I will be as surprised as if New York city turns out to
have been a media hoax.  (I am *not* asserting that either is
absolutely impossible.  But I've been to New York and I've spent years
studying particle interactions in bubble chambers, and you'll have to
come up with an awfully good story to convince me I've been had...)

If these gadgets *have* been working as described, then Special
Relativity correctly describes some very important properties of the
universe we live in.  We may someday find better ways of describing
these properties, but they won't go away, and we might as well give
them names.  One of them is the oblivious zone.

[83] by Arthur Protin (continuation of [81])

I will rule out neither faster than light travel nor time travel;
and I will not require them to be linked.
 
[84] by Jeff Prothero (response to [83])

Sure, and clocks may start running backwards and fairy godmothers may
fill the air.  Nobody is claiming otherwise, we're just trying to
figure out how to deal with the things we *know* about. Rod Serling
can fend for himself.

Unlearning "truth"s which fail can certainly be an uncomfortable
experience. :-)

[85] by Arthur Protin (continuation of [83])

Thus, I am quite willing to leave as easily undistinguishable that
which is concurrent and immediately interactable and that which is
concurrent and remote.  How I infer concurrence should not be part of
the language.

[86] by Jeff Prothero (response to [85])

Are you willing to leave as easily undistinguishable that which is
future and that which is past?  Can you present any argument other
than blind habit as to why the past/future distinction is more
deserving of a LittleWord than the present/oblivious-zone distinction?

(Some linguist shoot me down, please: is it Japanese which lacks an
explicit future tense?)

[87] by Jeff Prothero (continuation of [57])

Note that a world-wide optical network connected to sub-nanosecond
computers will shortly smash the traditional Netwonian assumption that
you can communicate so much faster than you compute that communication
can be treated as instantaneous: the local planet is 10^8
light-nanoseconds around.  The human cycle time being aproximately a
tenth of a second, this is analogous to holding a conversation in
which it takes sound (and light :-) a year to make it all the way
across the table.  (To complete the analogy, we may assume that
conversations last a million years and humans live many billions of
years... a year is still a long time to wait for the punchline! :-)
Dealing with the oblivious zone may become an everyday experience.
(Letters already cross in the mail, but in principle you can get
around that by phoning...)

The above example is still pretty Newtonian in most respects: all the
events are in the same inertial frame of reference, so our RNA logs
can still be synchronized sensibly: we can maintain the illusion of
"simultaneous" to some extent.  With a more complex scenario involving
observers on independently moving relativistic spacecraft, (or a
computer designed using optically coupled relativistic electrons
confined in suitable magnetic structures) we could remove this prop as
well.

Taking a different tack, if your sister radioported to Altair IV to
study compact-matter engineering and hasn't yet returned, you can't
really say that she *was*, *is*, or *will be* close to finals since
future and past tense clearly don't apply and present tense really
only makes sense (on human scales) for events local to your solar
system -- you really need to say that she *tense four* finals.  The
inappropriateness of the present tense would be intuitively clearer if
local stars, or many observers, had relativistic velocities...
perhaps it suffices to note that she is so far out in your oblivious
zone that for all you know she will turn out to have washed out of
engineering a year before finals and do-*tense four*ing a thesis on
the stylistic influence of DNA art on cetacean doppler music.

Recapping with considerably less precision than JimC would tolerate:

Newtonian tenses consist of:

   past
   present
   future

We can give physically meaningful definitions of these (which are
equivalent to the standard defintions of them whenever we can treat
the speed of light as being effectively infinite) like so:

An event is in our past if we could have received some sort of signal
from it (canonically, a photon) but not vice versa.

An event is in our future if it could receive some sort of signal from
us, but not vice versa.

An event is in our present if we can *exchange* signals with it.
Relativistically, this mostly makes sents if "we" are an event with a
non-zero duration in time (and, probably, space), so that the
"present" includes more than a single point of spacetime.  In the
Newtonian limit, this "present" includes all spacetime points
"simultaneous" with us, since light can get from us to any point and
back again, instantaneously.

Unfortunately, if we apply the above trichotomy to the universe we
actually live in, we wind up with a big left-over chunk of spacetime:
Those parts of the universe which could neither have sent us a signal,
nor have recieved one from us -- our "oblivious zone".  If we are
living on a timescale marked in picoseconds, then our oblivious
zone comes within microns of us.  If we are living on a timescale
marked in seconds, then our oblivious zone comes nearly as close as
the Moon.  If we are living on a timescale marked in decades, then
our oblivious zone creeps within a few stars of us.

This "oblivious zone" chunk of spacetime is just as significant as the
other three, and just as deserving of a name.  Just as we need a
future tense to refer to future events, a past tense to refer to past
events, and a present tense to refer to cotemporaneous events, so we
need a *tense four* to refer to events in our oblivious zone.

Is this at all worthwhile, or are we wearing a circular track in the
rug?

[88] by Bob LeChevalier (response to [57])

I see that this is an example where the choice of observer frame of reference
may lead to a different truth value for the claim 'A is before B', in short
'A is before B to observer C (in reference frame D?)'.  That the observer
and reference frame are necessary to the truth value does not negate that 
the statement is either true or false given the observer and reference frame.
Or at least I don;t understand th example if it is indeterminite given such
knowledge.

[89] by Jim Carter (response to [88])

Around a specified reference event the null (lightwave) cone divides the
universe into past, future, and a large zone between them with can only
be reached by faster-than-light travel and which hence can neither
influence nor be influenced by the reference event (if we lack FTL). 
It is very reasonable to me to call this the "present".  The null cone
is independent of coordinates so the "past" and "future" need no
reference frame specified.  However, by proper choice of coordinates
the time of any "present" event can be forced before or after the
reference, but that time difference has no physical significance.  In
such a situation it would be better to say that "A is relativistically
disconnected from B"; of course the gismu now assigned to "present" or
"simultaneous" would be used.  

In daily life the thickness (duration) of the present is negligible,
but not in astronomy, computing, satellite communication, etc.  For
example, the maximum diameter of an Ethernet system (1500 meters) is
controlled by how much delay can be tolerated for the thickness of
the present, so transmitters which start up "simultaneously" (+-
thickness) become aware that the other transmitter has spoiled their
signal before assuming that the packet has gone out.  

Jeff has some interesting ideas concerning events that extend over a
region of time and space.  However, given that Lojban tenses are
somewhat extendible, I think we can wait until star travel arrives to
deal with that problem :-)

[90] by Jeff Prothero (response to [88])

One of the themes I've been playing (more of a bagpipe drone :-) is
that, with a relativistically sensible tense set, "A is before B" is
observer-independent -- it will be true for one observer only if it is
true for all observers.  It is only the Newtonian tenses which have
the irritating property of being partly in the eye of the beholder.

(Irrelevant comment: I'm certainly not slamming Newton's accomplishments!
Anyone at the top of Gauss's list of great minds...)

[91] by Bob LeChevalier (continuation of [88])

A predicate language has to be fairly careful in its place structure policy,
and I think this is handled by ours.  For another, more down to earth example
the gismu 'bajra' (run) was once x1 runs from x2 to x3 via x4, modelled after
klama (go).  But then someone asked about running in circles and/or running
in place, where some of the places have NO true value, so we changed the
place structure  (it is now x1 runs on limbs x2, and if you want to go 
somewhere running, you make the lujvo  bajrykla (runningly-go) which analytic
ally gives the other place structure.

But why only 4 tenses?  You just argued that there might be 5 needed. 
And given some of the strage theories like superstring theory with its 8
dimensions, why not 8 tenses (assuming that tense is relevant to the others).

[92] by Jeff Prothero (response to [91])

I thought 10 was the most popular number ... ?  Anyhow, the "extra"
dimensions are rolled up too small to matter. How wide is the
universe?  Smaller than a proton -- if you travel through any of the
last six dimensions, har har!

[93] by Bob LeChevalier (continuation of [88])

And what if relativity is found to be just another successive approximation
to reality, as seems likely to me, at least?  What happens to tenses when
you have tachyons?

[94] by Jeff Prothero (response to [93])

Absolutely no question of this -- (classical) quantum mechanics
doesn't handle gravity, and relativity says zip about quantum effects.
Contemporary science has virtually no idea what happens when they
interact... albeit lots of interesting leads.  (It bothers me that
none of the quantum approaches seem to have a good explanation for the
Equivalence between inertial and gravitational mass, which Newton went
to great pains to verify, and which General Relativity explains
beautifully. Either I misunderstand, or a beautiful insight is being
ignored...?! Superstrings and twistors may escape this criticism, as
they have a strong geometric component...)

But, as Asimov points out, new science doesn't invalidate old
insights, it just adds more decimal places.  Any child quickly
discovers that the ground isn't flat on the meter scale.  It has been
known for millenia that, moutains aside, the ground is flat to a
percent or two on a scale of tens of kilometers.  At least a few folks
have known for the last couple of millenia that the ground is
spherical to within a few percent on a scale of thousands of
kilometers.  None of these insights invalidates the previous insight.
Astronomy progresses, but the Sun keeps rising in the East (for us
nonpolar types :-).  Special Relativistic effects are just a real and
well established as gravity -- the oblivious zone is no more likely to
vanish than is gravity.

I find it hard to take tachyons seriously.  You can look through your
legs and see the world upside down.  You can look at electrons as
positrons going backwards in time. Yawn.  

In real-world terms, the speed of light *is* infinite, in many useful
senses.  For example, in the Newtonian model, as you get closer and
closer to infinite speed, you can travel further and further in a
second, without limit, and exactly at an infinite speed, you can get
anywhere instantly.  

In relativistic terms, this remains true, with "lightspeed"
substituted for "infinite speed": as you get closer and closer to
lightspeed, you can travel further and further in a second (as
measured by your wristwatch), with no limit whatsoever.  In the limit
of actually reaching lightspeed, you can travel anywhere in the
universe instantly, as measured by your wristwatch.  What could it
possibly mean to go faster than that? You really expect to see your
wristwatch run backwards??

The intuitional problem comes when you travel "instantly" to another
galaxy, buy a candy bar, then hop "instantly" back to find the Sun has
burnt out.  Well, I *told* you "simultaneous" doesn't make sense and
that the present tense doesn't go that far. (Hotel Solaria: "You can
leave the present any time you like, but you can *never* check back
in!")  If you tune your language and tense system to emphasize that
"now" only includes suitably nearby things, this double hop won't
surprise you at all.  As long as you think in a Newtonian tense system
with a "now" that goes to infinity, you're going to keep finding that
the universe doesn't behave as your language does, and feel
indignantly that the universe is cheating!  *I* think that this is
(arguably) a wonderful example of a Whorfian construct at the heart of
the language getting such a strong mindlock on millions of people that
they can't imagine an alternative and honestly blame the universe for
not working like their language!

(For example, we say "The Megallanic Clouds *are* our closest
neighbors," implying that they are part of the present universe just
like New York and Iraq.  But they aren't, in as fundamental, physical
a sense as the past and future aren't part of the present -- as the
above double-hop shows.)

[95] by Bob LeChevalier (continuation of [93])

Lojban is already VERY avant garde among human languages in making time and
space tenses almost completely symmetric gramamtically.  We haven't left in
the time travel tenses, but there is an obvious spot in the grammar if they
are ever needed.  Otherwise, the structures for time and space are identical.
But separate in the default because MOST people don;t think in terms of time
and space in the same breath.  Lojban at least allows you to.

An essence of Lojban's design philosophy, carrying out one of JCB's ajor
ideas, is that Loglan attempts to "remove constraints on thought".

[96] by Jeff Prothero (response to [95])

Aye, and there's the rub:  To think and speak relativistically takes
four tenses, to think and speak in the Newtonian model takes only
three.  Loglan adamantly restricts you to three.  Repeat after me:
 You *will* think Newtonianly!
 You *will* think Newtonianly!
 You *will* think Newtonianly!
 ...

[97] by Bob LeChevalier (continuation of [95])

Wherever practical, we;ve designed choices into the language that seem to
remove a constraint - to make something obligatory in most langauges optional
in Lojban.  Our tense system is arguyably the most powerful such in human
langauge, though I occassionally find places where given infinite lexeme
size we might do better.  I challenge Jeff and anyone else to actually look
at the tense grammar and the lexeme members we have.  You'll blow your mind.
(In Lojban you can orbit a point in time, whatever that might mean.)  If
Jeff and others can come up with a way that breaks it and is needed for human
communication, though, then we should consider further change.  But the window
of time to do this is small, so start studying!!!

I reserve judgement about whether you are just daft, Jeff, or maybe ornery :-)
I also have no idea how trivial the question is.  We've spent far longer 
discussing more trivial points in the Loglan design before.  If a change to
the language design is really needed, it is not too trivial.
Because I don't know, it is NOT a waste of time.  The baseline grammar is up 
he flaws in the language before the textbook is done and the dictionary, and 
the language design is frozen, probably indefinitely for points like these.
I would urge people to study the grammar, and look for more significant
questions (they are certainly there - because we've made undocumented decisions
that no one has discussed publicly).

[98] by Jack Waugh (response to [91])

The four tenses sound good to me as a reflection of our understanding
today of time, communication, etc.  The fifth tense is not so
fundamental; it's just the start of a countably infinite series:
The two things are close enough in space and long enough in
duration so communication at light speed can get from A to B
and back to A, the same thing and back to B, the same thing and
back to A, and so on.

The response someone gave about what about if our understanding
of physics changes again, ties in with my ideas of another
letter today, on having two engineered grammars.  If tense is
only expressed with predication in the deep grammar, it will
be poised to grow with physics.  You can put in ideas like
the fourth tense without worrying about Zipf, because for
the deep grammar you don't care.  The shallow grammar
would abbreviate expressions that need frequent expression,
such as the tenses that are used in everyday life.

[99] by Robert Chassell (response to [91])

I'm not yet convinced lojban needs more tenses/spatials/relatials than it has,
but will observe (`will' as in its original sense, of `want/wish to'.... 
'observe' used metonymically ...) that an acquaintance is working on a 
computer with nanosecond switching times.  His circuit boards are 18 inches
across and speed of light delays are a design/debugging issue.  

Protin's imaginary debugging session in the example in
his recent message is not fantastical.

[100] by Bob LeChevalier (response to [1])

I just spotted something in Jeff P's earlier writing on relativistic tenses
that raised red flags (Wed 2/13 16:55)

>Time, fundamentally, is about causality.  The most important
>implications of "A happened before B" are that A may have caused B,
>and that B could not possibly have caused A.
>
>In the Newtonian approximation to reality, the speed of light is
>infinite, interactions can happen instantly over arbitrarily large
>distances, and a neat trichotomy holds: given any two events A
>and B, either:
> 1  A happened before B, and potentially influenced B.
> 2  A happened after  B, and potentially was influenced by B.
> 3  A was simultaneous with B.
>Newtonian physics allows no other possibility.

I assert that the first sentence is an assumption that Jeff is making
that is not necessarily true, and that the conclusions that result from it
are therefore invalid.

[101] by Jeff Prothero (response to [100])

You're touching on some fundamental issues, Bob, but I think you're
reading them backwards...

I'm certainly arguing from a set of propositions which I presume
we hold in common. (There is no alternative except violence. :-)
Among the propositions which I'm assuming we agree on are:

*  I'm assuming that Loglan is intended to be useful for humans
   who want to discuss events in the physical world.  If your
   stance is that Loglan is purely for discussing logical
   propositions, then we indeed are arguing at cross-purposes.
   In that latter case, the inclusion of tenses in Loglan is
   very hard to understand.  *Nothing* in pure logic requires tenses.

*  I'm assuming that we agree that there is a pragmatically
   significant difference between time and space.  I have done
   a crtain amount of pondering on this, and it seems clear to
   me that the sorts of properties and symmetries we observe
   along the spatial axes are indeed quite different from those
   we observe along the time axis.  I summarize these differences
   in a nutshell by observing that causality operates along the
   time axis, but not along the spatial axes.

I'm certainly willing to argue in support of either of the above
propositions!

[102] by Bob LeChevalier (continuation of [100])

Specifically, the assertion of relativity seems to
differ - that "A happened before B" does NOT imply that A caused B if B is
not in the light cone of A.  The alternative is to assume that the
definition of "happened before" is something in particular that justifies
the implication - in short, assuming the conclusion.

[103] by Jeff Prothero (response to [102])

When *using* a language, logic or theory, we use given statements to
determine conclusions about the world.

When *designing* a language, logic or theory, we work from known
properties of the world to determine what mechanisms we should
include.

It is perfectly true that the direction of synthesis is opposite in
the two cases.  This does not imply that either process is invalid.

In designing our tenses, we are partitioning spacetime up into a few
(three or four, in this discussion) mutually exclusive sets.  There
are an infinite number of ways of doing this. (Not just a coutably
infinite, or even Aleph-1 infinite number, either!)  We do *not*
select a partition at random!  We select that partition which we think
will be most useful to our audience. 

How do we do that?  Obviously, we cannot run a program to try all
possibilities and return the one that maximizes utility.  So we
set up some set of criteria, and construct a partitioning scheme
according to them.

The criteria I've focussed on are:

(1) It would be nice to reduce to the traditional tenses in the
    Newtonian limit.

(2) It would be nice if the auditor could make deductions from
    statements using the tenses without having to know anything
    about the person who made the statement.  (This is certainly
    entirely in the spirit of pure logic, and comes close to
    capturing what we mean when we say "formal".)

(3) It would be nice if the auditor could make deductions about
    potential causality on the basis of statements made using
    our tense set.

Given the above three design criteria, I believe (and, to a great
extent, science has demonstrated that) my proposed tense system is a
best solution, and nearly a unique solution.  (If jimc were to provide
his design criteria for a relativistic tense system, I think they
would be substantially the same, and he and I treat the "present" in
nearly diametrially opposite fashions, so there's at least some room
for disagreement within the framework.)

Is there a better set of design criteria, or does my design fail
to meet them?  Obviously, a concrete exposition of a better set of
criteria is more constructive than a simple rejection of the above
set.

So: Yes, my tense definitions are *deliberately* designed to allow the
auditor to conclude from "A before B" that "A potentially influenced
B".  Being able to draw new deductions from a given set of
propositions is entirely characteristic of logic, and enabling this is
an entirely proper design goal for a logical language.  The real
question is, why should we give up this capability by adopting a
naive-Newtonian tense system?  What do we gain in return for our loss?
The answer seems to be that we are giving up logical deductive powers
to avoid using one cmavo slot in LittleWord space.  A poor bargain,
in my opinion -- like giving up logical negation to free up a cmavo.

[104] by Bob LeChevalier (continuation of [102]) 

It is possible that in English, time and causality have some deep structural
relationship that justifies Jeff's assumption.  But Lojban, as a logical
language, need not, and indeed I believe does not.

[105] by Jeff Prothero (response to [104])

The link between time and causality is in the physics of the world we
live in, not in the language we speak.  English has developed among a
population of people who care, on a continuing and everyday basis,
about things like catching trains, planting crops, and not getting
killed... concerns which require accurate communication of information
about time and space.  I've been presuming that Loglan will be used by
people with similar interests, but there is of course nothing in pure
logic which justifies such a presumption.  

(If you *really* believe that there is no logical connection between
time and causality in Loglan, then you must believe that there is no
way for a Loglanist, told a that "A happened after B", to conclude
that "A did not cause B".  Are you willing to commit to this stance?
If so, Loglan detectives are going to suffer!  If not, Loglan's
reasoning rule's *do* link time and causality, since the first
statement concerns only time, and the second only causality.)

From the point of view of pure logic, we can drop the tense system
altogether, or use the "future tense" to denote events happening on
odd-numbered days, the "past tense" to denote events happening on
even-numbered days, and the "present tense" to mark events which never
happened at all.  Logic would be perfectly happy with this
arrangement.  Would *you* be perfectly happy with this arrangement?
I think you would *not* (whether you will explicitly admit it
or not :-) because such a tense system would not be tuned to conveying
the information you are most likely to want to convey.

I think that if you have the insight and honesty to explicitly dissect
out *why* the classical Newtonian tense system you favor is more
useful than the even-odd tense system above, you will find that it is
because the classical Newtonian tenses have associated with them
powerful reasoning rules which allow your listener to form additional
valid and useful conclusions from your statements, and that these
conclusions are centrally concerned with potential causality.  For
example, scholars give primary credit to the *first* publication of a
fact or theory.  Logic would be as happy with the *last* or *best*,
but from the standpoint of potential causality, the *first*
publication is uniquely positioned.  I don't think this consideration
is *central* to priority disputes and most other discussions driven by
tense-carried information.

Since neither you nor anyone else has tried to formulate a formal
semantics for Loglan, I can't examine the reasoning rules *you* have
postulated for the tenses, and am necessarily reduced to offering my
own and guessing at yours... letting you arbitrarily dismiss mine and
object that my guesses are mere straw men.  My formulation of tenses
is designed to allow the listener to make valid deductions about
potential causality.  In the relativistic regime, the naive Newtonian
formulation does not allow this, and the four-tense relativistic
formulation does.  What does the naive Newtonian formulation offer in
return?  Or is conveying useful information to the listener completely
irrelevant to Loglan's mission?

[106] by Guy Steele (response to [104])

If you ask me to take an absolute, black-or-white position on time
and causality, then I would have to conclude that, yes, every event
has a causal influence on every event in its future light cone and is
causally influenced by every event in its past light cone.  But most
of those causal influences are negligible.  They are of order 10^-40
or less relative to other causal influences.  It would be wrong of me
to claim that astrology is bogus because the stars cannot influence
us; the very fact that we are aware of their existence is evidence
that they can influence us.  But I might claim that their influence
is negligible with respect to the phenomena with which astrology
attempts to deal.  So we must ask ourselves whether Lojban is to make
itself pragmatically useless by insisting on 100% pedantic accuracy,
or whether we ignore negligible relationships most of the time.
(Consider, by comparison, the English idion "the lion's share".  If
you have read Aesop, you know that the lion's share is, in fact, the
whole thing, 100%.  But most people use it to mean 80% to 99%; in
other# words, the lion's share (sense #2) of references to "the lion's
share"# use sense #2, even though it is not true that the lion's share
(sense# #1) of references to "the lion's share" use sense #2.  This is
because a colorful idiom for "99%" is much more useful than one for "100%".

[107] by Bob LeChevalier (response to [105])

My essential disagreement with you is that I do not admit that Lojban's
tense system is inherently Newtonian.  What it infers is subjective time
which is relative to a single observer.  Tense (unless explicitly noted)
should not be referring to God's omnipresent and objective (presumably)
point of view.  I'll append a posting to this I was going to post separately
that elaborates.  Re the association between time and causality: you and I
may assume it, but is it a mandatory epistemological/metaphysical assumption
that must be embedded in the language.  Noting that Pynchon's Gravity's
Rainbow is  based on the possibility that causality is NOT time sequenced
(or rather that it is possible that an event in the future could cause one
in the past), it is at least conceovable that the opposite holds.

John Cowan, Athelstan, Nora and I (and eventually pc by phone) got
together this weekend.  Among the topics was the lively (40 printed pages
so far) discussion of relativistic tenses in Lojban.  Some highlights:

Tense is essentially subjective time; it is not a claim about absolute time
from God's (omnipresent) point of view, because the speaker may not know
what God's point of view is.  This is essentially the same as what Arthur
Protin was saying about speaker orientation.  To a peasant with no
astronomical training, the stars are 'just above the sky' and most
certainly are there in the present tense, regardless of what physics says.
In Lojban, all truth and knowledge has a place for epistemology.

The language allows the creation of predicates that specifically address
measured or objective time, and this is the preferred means for relativistic
references.

It was agreed that tense makes no claim about causality, as I've stated.
Indeed, I might generalize this to say that in Lojban, as a new language
with a bunch of only nebulously defined words so far, you can make NO
>assumptions< that predicate A necessarily implies predicate B merely
on the face of the definition, unless the definition explicitly includes
that implication.

There was consensus that if, in the future, a case was made to justify a
relativistic tense of the sort proposed, it is merely the addition of a
cmavo to lexeme PU.  The interaction of this tense with the rest of the
tense structure is of course unknown as to semantics.

pc, whose specialty is tense logic, listened to our summary of the
discussion, and indicated that he believed it to be a non-problem.  He
gave a reference that will send jimc and Jeff P. back to the research
books.  Apparently, a researcher investigated the problem of relativistic
tense thoroughly in the 1960's.  His PhD thesis is considered the
landmark study in the area, and has not been significantly challenged.
The conclusion was that no special modifications needed to be made
because of relativity.

pc did not have the exact reference, but the following should be
enough for jimc (who is at UCLA) to find, and probably Jeff as well.

  Nino B. Cochiarella is the author, date approx. 1967
  UCLA thesis
  title something like "A System of Relativistic Tenses in Tense Logic"

As to a formal semantics, our consensus was that there is no consensus on a
semantic theory and that to adopt one would be a strong metaphysical bias.
pc pointed out that it would hardly solve the problems that Jack Waugh is
concerned about, anyway.  We would need a formal theory of pragmatics for
that, and pragmatics theory is much less advanced than semantics.
(Pragmatics includes such issues as jimc's problems of anaphora resolution.)

[108] by Robert Chassell (response to [105])

Consider:

    `The flower grew and then I watered it.'

``Obviously'' (in some world views), the flower grew _because_ it
expected me to water it.  Newtonian and Einsteinian views of physical
causality are culturally recent innovations.  (Interesting and
valuable innovations, in my opinion, but that is beside the point.)

Loglan has words for different kinds of causality: physical,
entailment from rules of game, moral, and so on.  The tense system
should not presume one or other is primary.

If so, Loglan detectives are going to suffer!  Many detective stories use a
`means, motive, opportunity' format for figuring `who dun' it---Aristotle's
"final cause" is the type of cause considered, as well the "material cause"
and "efficient cause".

Loglan should be able to handle a world view that presumes that all
entities in the universe, including rocks, are sentient in some manner
and that all occurrences may be explained in a telelogical manner, in
which the causal events occur _after_ the caused event.

Of course, like the presumption that events described in a paragraph
are occuring at and after the time suggested by the last tense marker,
speakers will by default presume one or other causality system.

Once you stretch beyond the contemporary tense system of the
culturally recent past, tenses in English begin to look weird:

    The use of a form of desire to indicate the future, as in 
    "I _will_ go to the store." 

    The use of spatial movement to indicate the future, as in
    "He is going to build a house."

Perhaps equally useful weirdnesses will arise in lojban.

The idea that a littleword should be reserved for some kind of
relational other than the conventional contemporary tenses is
interesting.  I am not sure whether the relational should be a tense
or a spatial or something else.  On the other hand, the number of
littlewords is very small.  As a practical matter, I think that a
block of littlewords should be reserved for future use, without
specifying what they are for.  


[109] by Bob LeChevalier (continuation of [104])

Lojban's system of expressing causality is independent of its system of
expressing time.  On the other hand, the system of expressing time and space
are highly interactive.  I justify my claim by looking at another area where
causality overlaps a distinct feature in English - that of implication in
logic.

[110] by Doug Landauer (response to [109])

I don't think Jeff is suggesting either that
	A before B		implies		A caused/influenced B
or that there should be any dependence between causality and tense.
It seems to me that all he was saying is that it would make sense to
make the set of tenses more complete; and that it would be a bonus, in
a practical sense, that completing the set of tenses from the three
newtonian ones to the four relativistic ones would allow us to say
	A "tense-four" B	implies		A cannot have caused B
and
	A before B		implies		A *might* have caused B.
As you can probably guess, I agree with him that this addition makes sense.


In English, when we say "If A then B" we often mean that "A causes B".
The example used in Loglan discussions historically is "If you water it,
it will grow."  Those same historical discussions (most recently in one
of the Volume 5 issues of The Loglanist - I believe Richard Kennaway was
one of the authors, and is on this list - care to comment, Richard?),
clearly indicate that in Lojban, the conditional is a purely predicate
logic one, and that causality is not part of the definition.  Elaborate
efforts were then made to allow parallel and independent assertion of
causality in a similar and parallel structure to the logical connective
form to specifically delineate the semantic opposition between the two
concepts.  We retained this distinction in Lojban.  Thus, for our example:

ganai do dunda loi djacu ko'a gi ko'a banro       the conditional
Either it is not the case that you water it or it grows.
If you water it, then it grows.

ri'agi do dunda loi djacu ko'a gi ko'a banro      the causal
Because you water it, it grows.

which both use lexer_G_935 of the lexer grammar to form the first compound.

A SIMILAR OPPOSITION EXISTS BETWEEN TIME TENSE, SPACE TENSE, AND CAUSALITY.

Indeed it can even appear in the same sentence form:

bagi do dunda loi djacu ko'a gi ko'a banro        the tensal (?)
After you water it, it grows.

or

pugi do dunda loi djacu ko'a gi ko'a banro
After you water it, it grows.

or

cagi do dunda loi djacu ko'a gi ko'a banro
Simultaneous with your watering it, it grows.

or even for space tense

vigi do dunda loi djacu ko'a gi ko'a banro
Where you water it, it grows.

It turns out that causals are permitted anywhere else that tenses are
permitted, and indeed all sumti tcita of lexeme BAI and their
conversions and negations can occur anywhere that tenses are permitted.
The only constraint, which is dictated by the LALR1 grammar, is that you
cannot use a nonce sumti tcita using FIhO inside lexer portions of the
grammar, since a FIhO 'modal' (to use jimc's word) can carry a full
predicate with all of its sumti and other grammatical garbage into the
modal, requiring in effect the full grammar to be recursively embedded
in one lexeme.

(It is these same sumti tcita that allow specification of observer and
reference frame.  Enjoy:

sega'agi do dunda loi djacu ko'a gi ko'a banro
Observing your watering it, it grows.

ma'igi do dunda loi djacu ko'a gi ko'a banro
In the reference frame of your watering it, it grows.

Whatever these mean, they are not causals, or tenses, or implications.)

In Lojban, any predicate claim has an optional potentiality component.
The unmarked form is formally vague on whether there is a claim of
actuality or potentiality; pragmatic context determines the implied
tense in such a statement.  But there is at least a claim of
potentiality in the ellipsized tense.  I would maintain that discussion
of potentiality, which is carried in the grammar separately from the
tenses (although part of what might be called the overall tense
structure) is also independent of both causality and time/space tense in
Lojban.

If tense relationships in Lojban do not imply causality or potentiality,
is there still a need for these extra tenses?

[111] by Jeff Prothero (response to [109])

Your points, while valid, appear irrelevant to the immediate issue:

1 Why do we define the standard three Newtonain tenses the way we do?
  Put another way: sorts of deductions does this formulation support?

2 Given the above, how can we extend those definitions into the
  relativistic regime while doing minimum violence to them --
  while preserving their logical properties as much as possible,
  so that the sorts of deductions we are habitually accustomed
  to making will continue to be valid?

I think we're failing to communicate because you are taking the
naive-Newtonian tenses as God-given (that is, your interpretation of
"A before B" as inviolable, even in the relativistic regime), and
reasoning from there, whereas I am taking the logical properties of
the tenses as a given and deriving the logical spacetime partition as
a consequence.

"A before B" really means that "A is the in *past* part of spacetime
as partitioned at B".  Whether this is true in a particular case
depends on just what rules we use to slice up spacetime.  You are
arguing for a naive-Newtonian partition (apparently) because it saves
a LW, I am arguing for an alternate partition because it comes as
close as possible to letting people continue to use their familiar
Newtonian-regime tense-related reasoning patterns even in the
relativistic regime -- it preserves the logical properties of the
tenses, to the extent that it is possible to do so.

As long as we refuse to explicitly acknowledge that there *are* formal
properties of tenses that people rely on, we can treat the loss of
those properties as a negligible issue.  When and if we try to specify
the formal semantics of Loglan, keeping the formal semantics small and
regular will loom just as important as keeping the formal grammar
small and regular, and adding one extra LW will seem a small price to
pay for perhaps halving the formal semantic description of the tenses.

[112] by Arthur Protin (response to [111])

Since long before Newton, the present tense was used to describe things
that are going on now, that is, things you can not possibly see the
beginning of but may if you are "fast" enough see the end of.  Yes, I
am sure that you could take long enough to get there that the Mississippi
would no longer be flowing, all its water stolen by the Achefallia
(I forget how it is spelled).  Back when fast was horse back, "now" still
meant that you might have to be  infinitely fast to observe it.  The
notion of causality and interaction were IMHO far less significant in the
understanding of "now" than was the concept of linear time.

"Now" will still mean what it did even for events outside of the zone
of interactability.  Let me offer the following SF situtation as an example.

I set up my banks and banks of lasers years ago and fired them such that
now even as we speak their moon is exploding.

(and if you were omnipresent you would be able to both observe my statement
and its truth.)

The event light years away is going on now, I caused it years ago and I
will observe the feedback of the event in years ahead.

Any good change to "now" has to deal with a non absolute time.  It is
not important to the notion of "now" that the senses may witness the effects
of two events in the wrong order, just that there is AN order.  If time is
not monotonic,  then the notion of "now" will have to be revised.  As
long as time is monatonic, then there is one true ordering of the events
and while we have only a limited vantage point to observe the reality,
we are omnipresent in  our model of that reality.  "Now" means concurrent
with the speacker in that absolute time.

I can't find any better way of dealing with this (as of now).

[113] by Arthur Protin (response to an unavailable message by Jeff Prothero)

You appear to misunderstand me.  I was in no way suggesting that those
devices did not exist and work as described.  I am saying that as Newtonian
physics described some of the available experiences but proved to be
a special case of relativity, so relativity may be only a special case of
a more complete and more correct model.  One that may add to or modify the
limitations demanded by relativity.  The speed of light may not be the
true limit.

[114] by Jeff Prothero (response to [113])

Certainly Newton, who initially appeared to have virtually exhausted
physics, in fact left many things to be said, and later physicists have
gone on to say some of them.  But Newton's fundamental laws and insights stand
as firm today as the day he wrote them.  (On the physics front, at least.
His alchemical and theological positions don't seem to have fared as well.)

It is not fair to claim that "The speed of light may not be the true limit",
implying that this would be no more surprising than the insights
Special Relativity added to those of Newton's Principia.  The role
of the speed of light is as central to Special Relativity as universal
gravitation is to Principia.  A *fair* comparison comparison would be:

  Just as it may turn out that gravititational attraction is *not*
  proportional to the masses and inversely proportional to the square of
  the distance, so it may turn out that the speed of light is not a
  fundamental limit.

And, in fact, General Relativity (and we're now in *much* deeper
waters than Special Relativity) *did* replace Newton's Principia as
the reigning description of gravity.  Gravitational attraction no
longer propagates instantly, but is limited by the speed of light
(funny how it keeps popping up, no?), and the attraction doesn't have
to be *quite* inverse-square, because space turns out not to be
*quite* flat.

(From General Relativity, we now know that the angles of any physical
terrestrial triangle do *not* add to exactly 180 degrees, and that any
physical parallel lines near Earth do *not* maintain a constant
distance between them.  But Euclid's geometry still serves us as well
in practice as it did him, two millenia ago -- it is true to a very
high degree of accuracy, as the Greeks discovered and nobody since has
disputed.  I haven't heard any proposals for striking the word
"parallel" from English *or* Loglan, whatever its status in Nature's
geometry...)

BUT: General Relativity or no, you can still plot the orbit of a
spacecraft just fine using Newton's laws, just as Newton plotted the
orbits of comets -- and in fact, NASA does so.  Newton's laws are
still taught to students and used by engineers on a daily basis, all
over the planet.  Nothing in General Relativity would make Newton hang
his head in shame, feeling that he had thoroughly screwed up.  

(For example, he explictly never claimed to know if his gravitational
force was real, or how it might be propagated, and would undoubtedly
be delighted to hear that progress had been made on this front.
Newton, the alchemical heretic, was *amazingly* prescient.  His later
writings give light a wavelike periodicity to explain diffraction, and
a spatial assymetry to explain polarization.  In retrospect, one is
almost surprised that he didn't formulate a good wave theory of light,
saving the rest of us a couple of centuries... just imagine if his
alchemical obsession had led him to play with batteries, electricity
and magnets!)

Equally with Special Relativity:  The fundamental role of the speed of
light in our universe is beyond doubt, but one can certainly pick
nits, and in fact can continue to learn new physics from examining the
lightspeed limit in more detail:

Light travels more slowly in dense media than in vacuum, so we can and
do observe particles travelling faster than light in liquids and
solids, emitting characteristic Cherenkov radiation -- a photonic
"sonic boom".

Recent theoretical calculations suggest light travels something like 1
+ 10E-19 (?) times faster when within a micron (?) of a charged flat
plate and travelling perpendicular to it, due to the altered nature of
the vacuum.

Photons may yet turn out to have nonzero rest mass, hence travel
slightly slower than the relativistic "speed of light".  (In which
case we will probably start calling it the "speed of neutrinos", or if
*they* have rest mass, "the speed of gravity".  We still have almost
no idea why particles have mass, much less why they have the
particular masses they do...)

You can send light the long way around the planet and take a short
cut, arriving "faster than light".  Clever, no?

It is conceivable that the universe has a more complex topology than
normally supposed, providing short-cuts ("wormholes") for some paths
-- again, if you send light the long way and take the shortcut
yourself, you may be able to arrive "faster than light".  (You still
couldn't get through the wormhole faster than light could, and
theoretical attempts to construct wormholes result in conditions so
extreme -- tidal effects that disrupt atoms, &tc -- that nothing *but*
a photon seems likely to make it through. The wormholes also tend to
collapse after a microsecond or so. Details.)

None of the above has physicists doubting the essential validity of
Special Relativity, or the reality of the effects it describes.

[115] by Arthur Protin (response to [114])

    I am beginning to believe that we have become distracted by many
issues of relativity.  Even if relativity is replaced by a more complete
theory that includes a non-null explanation, the practical
effects predicted by his work will still be observed.

Einstein never said (anything like) "consider this deep structure for the
universe and see how that prevents us from measuring the speed of light
relative to its medium (the ether)" rather he said (something like) "let us
accept the observation that we can not measure the speed of light relative
to its medium and see what the mathematical implication of that failure is") 
                                                          

[116] by Jeff Prothero (continuation of [114]) 

The utility of the proposed "fourth tense", like the utility of our
particle accelerators, doesn't depend on the eternal acceptance of
Special Relativity as the best available description of phenomena like
the oblivious zone, just on the *existence* of those phenomena:

* To destroy the practical utility of Newton's laws, you have to
  eliminate something like gravity;

* To destroy the practical utility of Einstein's formulation of
  Special Relativity, you have to eliminate something like the
  fundamental role of the lightspeed limit -- maybe E != mc^2, and
  isn't even close?

* To destroy the practical utility of the fourth tense, you have to
  eliminate the oblivious zone to which it refers.

Honestly, Art, I think you're primarily exhibiting humanity's age-old
xenophobic reaction to anything new.

[117] by Arthur Protin (response to [116])

While that may be close, it is probably the effects of having grown up
around a saleman, and I now can not be sold anything (I buy things only
because I decide to).  Even my physics instructors in relativity class,
and quantum mechanics could not get me to accept things just because they
said them nor even just because they were in print in the book.  If this
idea of yours is a good idea, then only the cold logic of its justification
will move me.


[118] by Jeff Prothero (continuation of [116])

Try a null test on yourself:

All your arguments about the possible collapse of relativistic physics
apply with equal force to Newtonian physics.  Thus, if your arguments
require the exclusion of the relativistic tenses from Loglan, they
also require the exclusion of the current Newtonian tenses from
Loglan.  If you'll provide a list of possible ways in which relativity
could collapse to render the relativistic-tense partition of spacetime
nonsensical, I will undertake to match you one-for-one with a list of
possible ways in which Newtonian physics could collapse to render the
Newtonian-tense partition of spacetime nonsensical.

In my opinion, there will be time enough to reform the Newtonian
component of the Loglan tense system when and if Newtonian physics
*does* collapse (as opposed to collecting marginalia), and there will
be time enough to reform the relativistc component of the Loglan tense
system (if we get one) when and if relativistic physics collapses.
Either eventuality would be an unprecedented disaster (to what would
you compare it -- the failure of astrology??), and trying to guard
against such a historically unprecedented disaster strikes me as being
much like worrying about the possible sudden breakup of the planet
while building your house.

Even defining what one *means* by the phrase "faster than light" is a
vexing problem, if one makes any attempt at all to make it consistent
with what we know about the universe.  For example: travelling faster
means covering a given distance in less elapsed time, by ordinary
intuition.  Travelling at lightspeed means covering any given distance
in *zero* subjective elapsed time.  Thus, travelling faster than light
presumably means covering a given distance in less than zero elapsed
time.  This is a difficult concept for me...  Again, any massive
observer boosted to lightspeed has infinite mass, as measured by a
naive observer.  A massive observer boosted to more than lightspeed
would thus presumably have more than infinite mass, which is again a
difficult concept.  Anyone travelling FTL will appear to be travelling
backwards in time to some observers.  If you really let timetravel in,
it is going to be very difficult to keep *any* conservation laws in
effect, no?  As I commented earlier, if time travel is possible, it is
quite difficult to explain why the extreme conditions of the Big Bang
didn't propel particles forward in time, to rain down throughout our
contemporary universe... everywhere one looks, things seem to be
coming unstrung. It's like adding the axiom "2+2=5" to arithmetic
and then frantically trying to patch all the problems that follow.

[119] by Arthur Protin (continuation of [113])

I have not had time to review the texts on relativity, and since
you claim to work with it, can you post any examples of situations that
require time to not be monotonic, and/or prohibit an absolute ordering
of events (in an inhabitable reference frame)?

[120] by Jeff Prothero (response to [119])

Let's correct a possible mis-impression: I'm not a physicist, I get
my physics from Isaac Asimov & Co just like most folks, and any *real*
physicists reading the list are probably snickering from time to time.
The undergraduate Computer Science program at the UW opened up after
my freshman year, and I never took another physics course.  I *did*
hold a no-brain student helper job at the UW, scanning photographs of
neutrino events in the Fermilab two-meter bubble chamber. The
accelerator wouldn't work without relativistic corrections, and some
of the particles I saw decaying have lifetimes so short that they
wouldn't have made it far enough to be observable in the absence of
relativistic effects.

Excluding from discussion things like the quantum foam of spacetime as
observed near the Heisenberg limit, where all bets are off (for
example, quantum fluctuations can create masses large enough to change
the topology of spacetime! -- this is where quantum effects and
General Relativity meet, a theoretical mess), and restricting ourself
to the scale and era Special Relativity is designed to describe:

Any pointlike observer can always assign an order to all observed
events, simply by time of observation.  Non-monotonic time is the sort
of paradoxical concept that makes no sense in a straightforward
interpretation.  Time is measured by the successive states of some
physical system.  In the absence of time travel (or faster-than-light
travel), no clock will see another clock running backward, although
other clocks may appear to tick too slowly or too quickly.

[121] by Arthur Protin (response to [120])

Excellent!  Given that the universe is such a physical system, then
there should be an "absolute" reference for time (as marked off by the
states of the universe).  Even if we can not experience it directly,
we can assume its existance and attempt to define the correction
from our experience of time to that conceptual absolute!!!

[122] by Jeff Prothero (continuation of [120])

The problem is that the observed order of events often depends on the
observer, rather than the events.

[123] by Arthur Protin (response to [122])

Excellent!  Given that the universe is such a physical system, then
there should be an "absolute" reference for time (as marked off by the
states of the universe).  Even if we can not experience it directly,
we can assume its existance and attempt to define the correction
from our experience of time to that conceptual absolute!!!

[124] by Guy Steele (response to [123])

No, you have fallen into a trap here: the use of the word "state"
begs the question.  Implicit in the word "state" is the notion that
you are taking a "snapshot" of the entire universe *at the same point
in time everywhere*--but that is exactly the notion we are trying to
get a grip on!  Observers in different inertial frames will differ on
what constitutes a state.

[125] by Arthur Protin (response to [124])

Here you say that states are subjective, but the various subjective states
that you referred to are misdirection.

[126] by Guy Steele (response to [125])

Yes; I was sloppy in making a transition and I am sorry I confused
you.  In effect, I first cited the "intuitive" notion of state (a
snapshot) and pointed out that that this definition will not do
because it is observer-dependent.  I then proceeded, in the paragraph
beginning "What is a state?", to appeal to a more basic intuition
about state without recourse to the notion of global simultaneity
that is implicit in the word "snapshot"; then I used this more basic
intuition to *redefine* the word state so as to have an
observer-independent meaning.  Note that such states do not have
unique successors; they are only partially ordered, not totally
ordered.  State X is earlier than state Y if and only if there do not
exist two events x (in X) and y (in Y) such that x lies within the
future-light-cone of y.

[127] by Arthur Protin (continuation of [125])

There is either determinism (and we can keep our physics) or there isn't.
I will assume for at least this lifetime that there is determinism!

[128] by Guy Steele (continuation of [124])

What is a state?  It is the minimal amount of information needed to
imply future states (assuming determinism--here I ignore quantum
mechanics).  I argue, then, that to include in a single state
information about two causally related distinct space-time events
(locations) would be redundant, because some information about the
causee would be implied and could be omitted.  (This argument is not
air-tight, but perhaps could be made so.)  It follows that a state
ought to consist of information about a set of events that are not
causally related.  In other words, the set should be "space-like".
And this is in fact about the best we can say in a relativistic
theory, independent of any particular inertial frame: that any
"spacelike" slice through the four-dimensional universe may be
validly regarded as a state.  But another observer may regard
different parts of that slice as belonging to different states as he
observes them.

So, I am sorry, but the "successive states" model simply does not
necessarily define absolute global time.  It does define local time,
and that is consistent with relativity.

[129] by Jeff Prothero (response to [125])

The simple fact is that many things which seem "self-evident" are not,
and indeed a good number turn out not to be true.  (Part of the
delight of science!  Was it Eddington who said "The universe is not
only stranger than we imagine, it is stranger than we *can* imagine"?
But even if we cannot imagine Nature's wonders, we can discover them,
if we keep a sharp eye and open mind...)

It is "self-evident" that objects at rest tend to stay at rest, and
that objects in motion tend to come to rest.  Anyone playing with toy
boats in a bathtub can observe this.  It took the genius of Galileo
Galilei to realize that in fact there is no such thing as absolute
motion or rest -- motion can be measured only relative to some other
object(s) -- and that all inertial (unaccelerated) frames are
equivalent...  and it took Newton to codify this neatly.

It is "self-evident" that different observers can agree when two
events are simultaneous.  It took genius to realize that in fact the
simultaneity of two (spatially separated) events is dependent on the
observer's inertial frame of reference, that different observers in
different frames will see things differently, and that nature doesn't
favor any particular frame -- observations made in one frame are just
as "true" and valid as those made in any other.  (I'm not sure if
Einstein gets full credit for this insight... the Lorentz Transforms
were common currency, and there is general agreement that if Einstein
had not done Special Relativity, someone else would have done it
shortly.  Not to knock Einstein -- he has forgotten accomplishments
overshadowing the complete careers of most physicists...)

[130] by Arthur Protin (response to [129])

So the problem with our tenses is not talking about interactivity
(which you have succeeded in convincing me belongs in separate words)
and observing relativistic events but in communicating about events
with beings in a different inertial frame of reference (BIADIFOR).

[131] by Jeff Prothero (response to [130])

Useful communication between observers in different inertial frames is
certainly a central issue here!

[132] by Arthur Protin (continuation of [130])

When communicating with BIADIFOR, what tags will be useful for
translating between the DIFOR's?

[133] by Jeff Prothero (resopnse to [132])

There is a fundamental problem here: tenses convey *qualitative*
information.  When we use tenses to describe the relationship between
two events A,B to somebody else, we usually start with quantitative
information -- we normally have a pretty good notion of the distance
in time and space between the two events -- but we throw away most of
this information, and tell our auditor only which (of a small fixed
set: *before*, *during*, *after*...) conceptual pigeon-hole into which
the relationship fell. (In statistics, this is called "binning", and
Numerical Recipes recommends delaying it as long as possible, because
of the information loss.)

When we reduce our quantitative information about the relationship
between A,B and ourself to a simple tense, we throw away the
information needed to transform this description into a different
frame of reference.

I see only two ways around this problem:

(1) Include in our statement enough quantitative information about
    A,B and ourself for the auditor to transform into any desired
    inertial frame of reference.  In my opinion, this approach violates
    the spirit of the tense system, will be unacceptably verbose,
    and will require more arithmetic ability than most Loglanists
    are willing to acquire.  (I *like* the idea of a language which
    forces a little arithmetic on the user, but I don't think RLC
    would consider this to be in the spirit of Lojban. :-)

(2) As I have been advocating, use a set of pigeon-holes which
    don't *need* translation from one inertial frame of reference
    to another, because they use boundaries which are physical
    invariants, appearing the same to all observers:  the
    speed-of-light null-cones defined by one of the two events
    A,B.  If we define A's "future" to be those events lying
    within A's forward null-cone, and A's "past" to be those
    lying within A's backward null-cone, then all observers
    will agree on which events call for the past or future
    tense with respect to A, with no need for explicit
    transformations between different inertial frames of
    reference.

[134] by Arthur Protin (continuation of [132])

When communicating with BIADIFOR's becomes a common experience (at
least common enough that every body has met or spoken with some one
who has communicated with a BIADIFOR ("complete second hand
exposure")) it will be important that we have the tools to make those
translations easy.  That gives you some time to make a few false
starts (as obviously you have and will again) before you find a good
provision for it.

(If you still think that relativity and interactivity affect our
notion of "now", realize that the size (in four dimensions) of the
zone ("cone") of commonly agreed "now"ness is a function of the
difference between the velocity vectors of the two observers IFOR's!)

[135] by Jeff Prothero (response to [134])

Um, my formulation uses tenses to describe the relationship between a
pair of events, using *one* as the reference: "now" doesn't depend on
pairs of observers.  I think my relativistic "now" is as well defined
as the classical Newtonian "now", but won't claim to have found a
clearly ideal formulation.  ("Now" is admittedly a heuristic concept,
in the Newtonian *or* the relativistic formulation: "now" may cover an
interval of a billions years or a femtosecond, depending on the
subject at hand, with the auditor expected to guess correctly.)


[136] by Arthur Protin (continuation of [134])

My position is:
     1) The tenses are a crude reflection of the sequencing of events.

[137] by Jeff Prothero (response to [136])

They are a qualitative classification of possible relationships
between pairs of events.  Any qualitative presentation of quantitative
relationships may be legitimately termed "crude" :-)

[138] by Arthur Protin (continuation of [136])

    2) Since that sequencing is dependent on the inertial frame of
    reference (IFOR) of the observer with no known physical bias
    toward any one IFOR and no known way of removing the IFOR from
    the description, our language will remain in the speaker IFOR
    and may someday have hooks for ease of translation.

[139] by Jeff Prothero (response to [138])

The naive-Newtonian tense classification is dependent on the
observer's inertial frame of reference; the relativistic tense
classification is not -- thus, there *is* a "known way of removing the
IFOR from the description".

(Finding descriptions which are independent of any particular
coordinate system is a central theme of General relativity in
particular -- hence the focus on tensors -- and of physics generally.
At the risk of entering my zone of ignorance, we may define as
"geometric" precisely those properties which are invariant under
change of coordinate system.)

[140] by Arthur Protin (continuation of [138])

    3) Attempts to encode a speed of light (SOL) distance into the
    tenses is bogus because:
        a) The second point (the origin) for the measurement is not
            always obvious,

[141] by Jeff Prothero (response to [140])

Tense is used to describe pairs of events; One of the two serves as
the reference point (origin).  In compound tenses, the reader is asked
to guess at (or accept as arbitrary) one of the reference points; this
is true in both Newtonian and relativistic formulations.

[142] by Arthur Protin (continuation of [140])

        b) It is insufficient to aid the translation into other IFOR's,

[143] by Jeff Prothero (response to [142])

In a sense, but since such translation is not needed, the point is moot.

[144] by Arthur Protin (continuation of [142])

        c) It biases our thinking toward only those interactions that are
            direct and at SOL,

[145] by Jeff Prothero (response to [144])

It makes physically important distinctions, as opposed perhaps to
politically or theologically important distinctions: I think this
charge could be leveled against any tense system.

It tells us more about physically possible influences and interactions
than it tells about the other kinds of influences and interactions...
I suppose you could consider this a bias. :-)  Again, I think this
charge could be leveled against any tense system.

[146] by Arthur Protin (continuation of [125])

Only one state description will correctly imply the future.  That state
description is one of a sequence that is the absolute time reference.
That we can not measure, observe, or even correctly infer those
state descriptions will not deter me from the belief in their existence.
(A much more compelling proof on their non-existence will be required
to move me.)

[147] by Guy Steele (response to [146])

This is the flaw in your reasoning.  In a relativistic theory,
observer-independent states are not totally ordered, only partially
ordered.

[148] by Arthur Protin (continuation of [146])

Now what proof have you that we can not ever infer the "true and absolute"
event sequence for the events we do observe?

[149] by Jeff Prothero (response to [146])

None!  I don't even have a clear notion of what a "true and absolute"
sequence would be, given what we know about the universe, much less
how it might be established.  (A verifiable angel stating that "Allah
is the only God, Mohammed is his Prophet, and Mecca defines his
Inertial Frame," perhaps? :-) Can you suggest an experiment that would
establish such a frame?

I also don't have any proof that we will not one day discover that the
Earth is indeed at "true and absolute" rest, with the rest of the
universe rotating around it, just as Ptolemy supposed.

In both cases, it is clear that any such proof would be of very little
practical importance -- it is *abundantly* clear that in practice, the
Earth is most conveniently treated as being just one more planet, and
that in practice the local inertial reference frame is most
conveniently treated as being just one more typical inertial reference
frame.

For example, even if the Earth is "proved" to be at absolute rest, and
epicycles are "proved" to constitute the "true and absolute"
description of planetary motion (again, I haven't the foggiest notion
what such a proof would look like), NASA will go right on plotting
planetary spacecraft courses using the Copernican Sun-at-rest model
and Newton's laws, just because they are so much simpler and more
convenient -- they capture important truths, if not the only truths.

We're moving from science (descriptions of what we actually
observe) to religion (unshakable a priori beliefs) here.  No problem
-- religious beliefs are an accepted part of contemporary human
culture.  But is it really fair to ask Loglanists to use a tense
system based on phenomena which cannot be measured, observed, or
correctly inferred?  Would it not be more conservative to use a
relativistic tense system which depends only on what a normal person
can reasonably observe and deduce?  Should ESP be a pre-requisite to
learning Loglan?

[150] by Jeff Prothero (continuation of [122])

Any person or system interested in
surviving goes to great lengths to ensure that, as far as possible,
descriptions of events depend on the events themselves and not the
observer.  For example, the human visual system goes to great lengths
-- *cheats*, to put it bluntly -- to ensure that the percieved color
of an object depends on that object, and not on the color of the
illumination.  Similarly, the visual system goes to great lengths to
ensure that the perceived size of an object depends only on the
object, not on the distance between the object and the observer...  in
spite of the fact that the retinal image varies dramatically with
distance.  Mistaking a lion for a kitten can be fatal!

In similar fashion, most people, most of the time, attempt to make
their descriptions observer-independent.  We don't say "it was a
little wee train, shorter than my finger", just because it was
distant, nor do we say "it was a giant train, taller than the Space
Needle", just because we were next to it.  We *care* about the
difference between

 "There's a kitten in the bedroom, dear."
 "There's a lion in the bedroom, dear."

even if they kittens and lions sometimes look alike to some observers,
neglecting size and distance.  We expect the speaker to make a
good-faith effort to distinguish the two cases for us, and provide an
observer-independent description, not to provide unedited sense-data.

Thus, I have no doubt whatever that *real* people, dealing routinely
with *real* relativistic effects, would carefully distinguish between
real "A before B", when that relationship is true for *all* observers,
and mere apparent "A before B", when that relationship only indicates
observational happenstance. 

In summary:

 *  "Non-monotonic time" means almost nothing to me.  I could invent
    a meaning, but it wouldn't be very physical.

 *  Any *single* observer certainly *can* impose a subjective order
    on events.

 *  There is *no* absolute ordering of *all* events that all observers
    can impartially agree on.

[151] by Arthur Protin (response to [150])

Right, there will always be people that can not understand the laws that
affect their perception of the absolute reality, and as such will never
be able to correct their perception for those distortions.
Further, we may have to refine our models to more completely correct
our observations.


[152] by Jeff Prothero (continuation of [150])

 *  There are *some* events whose ordering all observers will
    agree on.  (These are the only cases where *I* would use
    the past and future tenses. I don't like to confuse apparent
    size with real size, and I don't like to confuse apparent
    ordering with real ordering.)

 *  For the remaining events, whose order lies purely in the
    eye of the observer, my proposal uses a fourth tense.
    JimC would (ick! :-) use the present tense for these, and the
    naive Newtonian tense approach (deceptively, in my
    opinion) uses the regular past and future tenses,
    lumping them in with the truly ordered events.
    ("Kittens, lions... hey, they're all cats, right?")

[153] by Arthur Protin (response to [152])

No! I am sorry, but we just showed that there is an absolute time
that can be infered and subjective observations can be translated
into that absolute reference!!!

[154] by Arthur Protin (continuation of [119])

One of those "other reactions": I don't like cluttering up a simple
notion of time with (what I believe is) an orthogonal notion of
interactablity.

[155] by Jeff Prothero (response to [154])

Riposte I: If you truly believe that "time" and "interactability" are
orthogonal, let us construct a table:

           A was long before B | A was during B | A was long after B
           --------------------+----------------+-------------------
A potentially                  |                |
influenced B                   |                |
           --------------------+----------------+-------------------
B potentially                  |                |
influenced A                   |                |
           --------------------+----------------+-------------------
A potentially                  |                |
interacted                     |                |
with B                         |                |
           --------------------+----------------+-------------------
A could not                    |                |
possibly have                  |                |
interacted with B              |                |
           --------------------+----------------+-------------------

Since the two axes carry orthogonal labellings, it should be possible
to find a pair of events A,B to go in each box.  Would you care to do
so?

Riposte II: It seems to *me* that the naive-Newtonian tenses, lumping
apparent event orders indiscriminately together with real event
orders, in fact have a more complex semantics than the straightforward
relativistic tenses.  In particular, the relativistic semantics
doesn't have to deal with the position and velocity of the speaker and
auditor, which the naive-Newtonian tenses do -- they having no other way
to distinguish apparent event orders from real event orders.

Riposte III: I've explained what *I* think the primary use of tense
information is -- to provide the auditor with the
information needed to make qualitative deductions about potential
causation and interaction.  For example, *I* regard the differences
between

 "There was a lion in the bedroom, dear."
 "There is a lion in the bedroom, dear."
 "There will be a lion in the bedroom, dear."

as crucially involving the *interactions* I can expect if I open the
bedroom door, and the results they might *cause*.  I place great value
upon being able to deduce such potential interactions, and would be
quite upset if the tense system were changed so as to reduce or
eliminate my ability to make such deductions.  (The linguistic
mechanisms we now use to convey tense obviously *could* be used to
convey much different information, such as the relative and/or
absolute social status of speaker and listener.  I suspect that there
are human languages which in essentially do this. Presumably Loglan
has good pragmatic reasons for making the choice it did...)

If you really *don't* see deduction of possible interactions as a
significant use of tense information, can you explain what information
you *do* get from the tenses in the above examples?  Can you specify
the design criteria for an ideal tense system?  Can you explain why
you care about tense at all?  What difference would it make in your
behavior and understanding if people completely stopped using tenses
in their communication with you?


[156] by Arthur Protin (response to [155])

I can easily fill out your table of time versus interactability
(this is becoming tedious), but you would complain that the interaction
was skewed. For example, the window shatters and an hour later the
boy brings home his new BB gun (obviously the BB traveled much faster
and more directly into the liviing room).

We need to be able to talk about the observation and the origin
independently.

When we say
  "There will be a lion in the bedroom, dear."
we are not limited to thniking about the interaction between the
lion and 
    the speaker,
    the listener,
    the bedroom,
but may be interested in any or all the things that may be in the
bedroom at any point of the lions stay as well as anything that
may be left or altered by the lion during  his stay that can interact
with anything that enters the bedroom after he has left.

Your energetic discussion of this topic has convinced me completely
that the requested "fourth" tense is bogus.  I want to be able to think
about time and the sequences of observations even when I can not
understand the causes and effects.  I do not want my language corrupted
by the forced linkage of time with arbitrary scopes of interactability.

[157] by Carl Burke (response to [156])

As I understand it, this debate which has been clogging my in-basket
is raging over whether or not to incorporate a fourth "time" cmavo,
for a set of "past, present, future, and *not applicable*".  The fourth
tense refers to (in a relativistic sense) events outside the perceptible
area of space-time; *fourth tense* events cannot possibly interact with
the speaker, at least in the location/interval in question.
Is this an adequate non-technical summation?

[158] by Jeff Prothero (response to [157])

An excellent summation, except I prefer "*other*" to "*not
applicable*".  "*Not applicable*" suggests that the statement cannot
sensibly be evaluated in context -- in computer terms, that it
deserves a "compile error" (or evaluates to "bottom".)  But it is not
senseless to discuss nonpresent, nonfuture, nonpast events in a
neighboring galaxy -- it just requires a fourth tense.

Certainly, we could concievably abuse a "*not applicable*" flag by
using it instead of a fourth tense, but this would be a semantic mess.
If someone asks what tense correctly describes the relationship
between pride and the color green, *then* I would invoke *not
applicable*, since pride and green are not spacetime events
describable by tenses.

[159] by Carl Burke (continuation of [157])

This seems to be a useful concept, but it would find applicability outside
the strictly physics-related relativity frame.  For example, a criminal's
alibi is an attempt to assert a *fourth tense* relation with the crime,
under the prevailing conditions.  Locked room mysteries, by limiting
the communications means, isolate the "world-line" of the crime; the
detective must identify the means by which communication occurred to reach
the goal of "solving the case".  In the limit, barring FTL communications
a la Bell's Theorem, you have the cases argued (ad nauseum) so far.

[160] by Jeff Prothero (response to [159])

(Bell's Theorem / EPR demonstrates "nonlocal effects", but does not
give us FTL communication. Let's not thrash *that* out here!  But I'll
nominate it for best example of the universe being stranger than we
could have imagined...)

Certainly, a literal tense-four relationship to a crime would be an
excellent alibi -- for example, showing that you were born in the
oblivious zone of the crime would be as good as showing you were born
after the crime.  But a locked room doesn't mean much to a physicist:
neutrinos, gravity waves etc freely carry information into and out of
human rooms. (A black hole, now, ...)  Such a use of the fourth tense
would be very loose and analogical.  In English, at least, I have the
impression that such loose use of the tenses themselves is rare to the
point of non-existence, although the corresponding predicates are fair
game.

[161] by Carl Burke (continuation of [159])

This would appear to be every bit as useful as the addition of a third
value to traditional Boolean logic: True, False, *unask the question*.

[162] by Jeff Prothero (response to [161])

I think there are actually a number of candidates for a "third"
logical value: "unknown" springs to mind, plus "true but unprovable"
and "orthogonal -- can be assumed true *or* false without danger of
contradiction or inconsistency". I suspect pc could provide a dozen
more.  Want to make up a list and run them by RLC?  They deserve at
least compound predicates, I should think.

[163] by Carl Burke (continuation of [161])

The third value holds where neither a true or false value is applicable;
the traditional question is "When did you stop beating your wife?"
I seem to recall that there is another way to assert this third truth
value in Lojban, but I do not recall the method.  If this is available,
then there is no need (other than shorthand convenience) for the fourth
tense; you merely state that *neither true nor false* *actor* *relation*
*arguments* ***at all times/places, if you must specify***

[164] by Jeff Prothero (response to [163])

The entire Lojban tense/locator system is (as RLC likes to remind us)
a shorthand convenience for things we could say via (appropriate)
predicates, and one can certainly handle the fourth tense this way,
with or without a third logic value.  But you lose symmetry,
compactness and compound tenses involving the fourth tense -- you add
enough verbosity to practically guarantee that nobody will actually
use the fourth tense: "Omit needless words! Omit needless words! Omit
needless words!" (Will Strunk.)

[165] by Jeff Prothero (general summary)

RLC/lojbab wanted a simple rule for using the fourth tense on an
everyday basis.  I think I understand it well enough now to give
supply a twenty-five words or less rule (-:

   Whenever you find yourself about to use the present tense
   to refer to the fixed stars, use the fourth tense instead.

(21 words!)  Rationale: In everyday use, the "present" is usually less
than a year long, which puts the fixed stars clearly out in the
oblivious zone.  With the exception of the fixed stars, contemporary
daily life rarely if ever has occasion to refer to events in the
oblivious zone.
</pre>