null quantification: Difference between revisions

From Lojban
Jump to navigation Jump to search
mNo edit summary
 
mNo edit summary
Line 1: Line 1:


== LESSON 2: Relationships and places ==
''(originally from [[jbocre: tu'o|tu'o]] and relevant to discussions of [[jbocre: tu'o|tu'o]] and [[jbocre: lo'ei|lo'ei]] and similar creatures.)''


=== Names and relationships ===
I don't understand how the meaning of the bridi can be determined without quantifying over the underlying set, unless the set is one-membered. Perhaps the idea is simply that the quantifier is left unspecified so that it is glorked from context? That seems reasonable, but it's not equivalent to ''[[jbocre: lo'ei|lo'ei]]'' (e.g. ''mi nitcu tu'o tanxe'' is not equivalent to ''mi nitcu lo'ei tanxe''). --[[User:And Rosta|And Rosta]]


In Lesson 1 we looked at ''cmene'', Lojban names.
;: If the set is one-membered, then you are still quantifying over that one member, though perhaps a bit trivially. I think that the idea (as [[User:xorxes|xorxes]] has expounded) is that the quantifier is deleted, resulting in the intensional meaning of the selbri being added to the main bridi without extensionally quantifying over the underlying set. This '''is''' just what happens with ''[[jbocre: lo'ei|lo'ei]]'': ''mi nitcu lo'ei tanxe'' is defined as ''mi kairnitcu le ka ce'u tanxe''. There is '''no''' quantification over ''lo'i tanxe'' there, only a quantification over ''lo'i ka ce'u tanxe'', which, like every property, whether it obtains or not (or can possibly obtain), is inherently a singleton. -- [[jbocre: Adam|Adam]]


''cmene'' always label one particular thing. Just as in English, if I say "Mary",
If we think of 'sets' as groups instead (collectivities, = Lojban 'masses'), then we can just refer directly to the set/group without quantifiying over its membership. Regarding your analysis of ''tu'o'' and ''lo'ei'', I don't get it. ''lo'ei'' always seems to reduce to ''lo'' with a narrow scope within that of some implicit predicate. In other words, I think I have a rough idea of how ''lo'ei'' works, but I can't make sense of your and xorxes's analysis of its working. Maybe this discussion should move to the [[jbocre: lo'ei|lo'ei]] page, if you're arguing that ''tu'o'' is equivalent to ''lo'ei''. --[[User:And Rosta|And Rosta]]


I mean one particular person called Mary, no matter how many people
;: Mathematical sets cannot be used directly, and mathematical sets are what ''lo'i'' does. Also, in Lojban you cannot just use groups/collectivities/masses directly without quantification. For individual gadri (''lo/le/la'') you must pick out (quantify over) individuals from that group. For mass gadri (''loi/lei/lai'') you must pick out (quantify over) parts of that mass. xorxes's rewrite of ''broda lo'ei brode'' to ''kairbroda le ka ce'u du lo brode'' was an intermediate step, I think, in the ultimate goal of getting to ''kairbroda le ka ce'u brode''. As xorxes says, if you don't understand that, take ''sisku'', which is a primitive in standard Lojban: ''mi sisku le ka ce'u tanxe'' (=''mi buska lo'ei/tu'o tanxe'') doesn't involve any quantification over ''lo'i tanxe'', it merely uses the meaning of tanxe as expressed in the property ''le ka ce'u tanxe'', and adds it to the predicate ''sisku''. --[[jbocre: Adam|Adam]]


there are in the world called Mary, so in Lojban, ''meiris.'' can
I don't mean that ''lo'i broda'' is a group rather than a mathematical set. I mean that if we conceive of categories as *groups* of individuals, then we can either quantify over the membership, or refer to the group directly. (This is how ''lei/loi'' ought to work.) Regarding ''sisku'', as I've been saying in the discussions on Jboske, I don't understand ''sisku tu'o ka ce'u broda'' except as a way of expressing ''troci tu'o du'u co'e lo broda'', so for me the analogy with sisku does not help at all in explicating the notion of null quantification. I understand that ''lo'ei'' is supposed to "use the meaning of tanxe as expressed in the property ''le ka ce'u tanxe'', and adds it to the predicate ''sisku''", but I can't make sense of this. We haven't found any examples from English that don't reduce to ''lo'' (or ''loi'e''). --[[User:And Rosta|And Rosta]]
 
only refer to one person.  This means that ''cmene'' can never
 
stand for classes of things (like "person", "dog" or "computer") or for
 
relationships between things (like "loves", "gives" or "is inside").
 
Relationships are the key to Lojban, and words describing a relationship
 
are called ''selbri''.  A ''selbri'' is not a '''type''' of word (like a "verb" in English), it is
 
something that some types of  word  can '''do'''. Various types of word can act as ''selbri'', but ''cmene'', as we've seen, can't.
 
The main type of word used as a ''selbri'' is a ''gismu'', or root-word.  These are the building blocks of Lojban vocabulary.  ''gismu'' are easy to recognise, because they always have five letters, in the form
 
;CVCCV: or
 
;CCVCV: (C=consonant; V=vowel).
 
==== Exercise 1 ====
 
Which of the following Lojban words are:
 
* (a) ''gismu''
* (b) ''cmene''
 
* (c) neither?  Note: I've left out the full stops in the ''cmene''~--~that would make it too easy!
 
# ''lojban''
 
# ''dunda''
 
# ''ankaras''
 
# ''mi''
 
# ''cukta''
 
# ''prenu''
 
# ''blanu''
 
# ''ka'e''
 
# ''dublin''
 
# ''selbri''
 
Now we can recognise a ''gismu'', let's see what we can make it do.  ''dunda'' means "give", and as a ''selbri'' describes a relationship between a giver, something they give, and someone who receives it.  Let's say we have
 
three people, Maria, Claudia and Julia.  If we say
 
''la mari,as. dunda la .iulias. la klaudias.''
 
we mean that Maria gives Julia to Claudia~--~let's say Julia is a baby, as since the abolition of slavery, we don't normally give people as presents.  In English you can "give" someone in marriage, but that's
 
a culture-specific metaphor, and Lojban discourages that kind of thing~--~it's an example of ''malglico'' ("bloody
 
English"), transferring features of English into Lojban which don't
 
work.  If, on the other hand, we say
 
''la .iulias. dunda la mari,as. la klaudias.''
 
we mean that Maria is the baby, and Julia gives her to Claudia.  How do
 
we know this?  English uses the word "to" to indicate the receiver, and
 
in some other languages (like Latin or Turkish) the form of the words
 
themselves change.  In Lojban, as in logic, we have what is called
 
'''place-structure'''. Place-structure means that
 
;''dunda'' doesn't just mean "give", it means:x{SUB()}1{SUB} gives x{SUB()}2{SUB} to x{SUB()}3{SUB}
 
where "x" means someone or something.  Even if we just say ''dunda'' on its own, we still mean that someone gives
 
something to someone; we just aren't interested in (or we already know who or what.
 
We can say, then, that ''dunda'' has three "places".  We can think of places as slots which we can, if we want, fill with people, objects, events or whatever.  These places are called ''sumti'' in Lojban (easy to remember, as it sounds a bit like someone saying "something" and chewing off the end of the word).  Again, a ''sumti'' is not a '''type''' of word, it is something a word '''does'''.  The simplest Lojban sentence is a ''bridi'', i.e. a ''selbri'' and a bunch of ''sumti''. In other words,
 
''bridi'' = ''selbri'' + ''sumti ''
 
'''Note for logicians and computer programmers:''' for ''selbri'' read "function"; for ''sumti'' read "argument."
 
How many ''sumti'' can a ''selbri'' describe? The number depends on the place structure of the word we use for the ''selbri'' (there are ways of tagging on extra ''sumti'', which we'll cover in later lessons).  A ''gismu'' has a set number of places; as we've just seen, ''dunda'' has three.  The number of places varies from one to a staggering (and rare) five. Here are some examples.
 
==== One place ====
 
;''ninmu'':x{SUB()}1{SUB} is a woman (any female humanoid person, not necessarily adult)
 
;''blabi'':x{SUB()}1{SUB} is white / very light-coloured
 
;''cmila'':x{SUB()}1{SUB} laughs [[jbocre: not necessarily at someone or something~--~to include the object of the laughter you would use the ''lujvo'' (compound word) ''mi'afra''~--~x{SUB()}1{SUB} laughs at x{SUB()}2{SUB}, a slightly different concept]]
 
==== Two places ====
 
;''cipni'':x{SUB()}1{SUB} is a bird/avian/fowl of species x{SUB()}2{SUB}
 
;''vofli'':x{SUB()}1{SUB} flies [[jbocre: in air/atmosphere]] using lifting/propulsion means x{SUB()}2{SUB}
 
;''jungo'':x{SUB()}1{SUB} reflects Chinese [[jbocre: Mandarin, Cantonese, Wu, etc.|Mandarin, Cantonese, Wu, etc.]] culture/nationality/language in aspect x{SUB()}2{SUB}
 
;''junri'':x{SUB()}1{SUB} (person) is serious/earnest/has gravity about x{SUB()}2{SUB} (event/state/activity)
 
==== Three places ====
 
;''xamgu'':x{SUB()}1{SUB} is good/beneficial/acceptable for x{SUB()}2{SUB} by standard x{SUB()}3{SUB} [[jbocre: This is '''very''' Lojbanic~--~the English word "good" on its own is so vague as to be almost meaningless.  It is also slightly ''malglico'' to put a person in the x{SUB()}1{SUB} place, which is normally filled by an object, state or event~--~or moral good you would usually use ''vrude''~--~"virtuous"]]
 
;''pritu'':x{SUB()}1{SUB} is to the right of x{SUB()}2{SUB} facing x{SUB()}3{SUB} [[jbocre: remember all those times you have to ask "Is that my right or your right?" in English]]
 
;''cliva'':x{SUB()}1{SUB} leaves x{SUB()}2{SUB} for x{SUB()}3{SUB} by means x{SUB()}4{SUB}
 
;''kabri'':x{SUB()}1{SUB} is a cup/glass/tumbler/mug/vessel/[[jbocre: bowl|bowl]] containing contents x{SUB()}2{SUB}, and of material x{SUB()}3{SUB}
 
==== Four places ====
 
;''vecnu'':x{SUB()}1{SUB} [[jbocre: seller|seller]] sells/vends x{SUB()}2{SUB} [[jbocre: goods/service/commodity]] to buyer x{SUB()}3{SUB} for amount/cost/expense x{SUB()}4{SUB}
 
;''tivni [[jbocre: tiv|tiv]]'':television x{SUB()}1{SUB} [[jbocre: broadcaster|broadcaster]] televises programming x{SUB()}2{SUB} via media/channel x{SUB()}3{SUB} to television receiver x{SUB()}4{SUB}
 
==== Five places ====
 
;''klama'':x{SUB()}1{SUB} goes/comes to x{SUB()}2{SUB} from x{SUB()}3{SUB} via x{SUB()}4{SUB} by means x{SUB()}5{SUB}
 
;''cukta'':x{SUB()}1{SUB} is a book about subject/theme/story x{SUB()}2{SUB} by author x{SUB()}3{SUB} for audience x{SUB()}4{SUB} preserved in medium x{SUB()}5{SUB}
 
;''funva'':x{SUB()}1{SUB} translates x{SUB()}2{SUB} to language x{SUB()}3{SUB} from language x{SUB()}4{SUB} with translation-result x{SUB()}5{SUB}
 
=== Determining place structure ===
 
If all these places sound a bit daunting, don't worry~--~you don't have to memorise all of them (in fact nobody does).  There are a few cases where it's worth learning the place structure to avoid misunderstanding, but usually you can guess place structures using context and a few rules of thumb.  
 
# The first place is often the person or thing who ''does'' something or ''is'' something (in Lojban there is no difference between "doing" and "being").
 
# If there is someone or something that has something done to them he/she/it is usually in the second place.
 
# "to" places nearly always come before "from" places.
 
# Less-used places come towards the end.  These tend to be things like "by standard", "by means" or "made of". The general idea is that the places which are most likely to be filled come first.  You don't have to use all the available places, and any unfilled places at the end are simply missed out.
 
==== Exercise 2 ====
 
Try to guess the place structure of the following ''gismu''.  You probably won't get them all, but you should be able to guess the most important ones. Think of what '''needs''' to be in the sentence for it to make sense, then
 
add anything you think would be useful.  For example, with ''klama'', you need to know who's coming and going, and
 
although you could in theory say "Julie goes," it would be pretty meaningless if you didn't add where she goes to.  Where she starts her journey, the route she takes and what transport she uses are progressively less important, so they occupy the third, fourth and fifth places.
 
# ''karce'' - car
 
# ''nelci'' - like
 
# ''cmene'' - name
 
# ''sutra'' - fast
 
# ''crino'' - green
 
# ''sisti'' - stop, cease
 
# ''cmima'' - member
 
# ''barda'' - big
 
# ''cusku'' - say, express
 
# ''tavla'' - talk, chat
 
=== ''gismu'' as ''sumti'' ===
 
So far we've seen how a ''gismu'' can express a relationship between two or more ''cmene'', so we
 
can say things like
 
;''la bil. nelci la meilis.'':Bill likes Mei Li
 
But if we don't know her name, how can we say "Bill likes the woman"? If we say ''la bil. nelci la ninmu'', we mean that Bill likes someone whose name is "Woman".  What we say, in fact, is
 
''la bil. nelci le ninmu ''
 
What does ''le'' mean here?  We translated it into English as "the", but that isn't quite it.  The best way to think of it
 
is "the thing(s) I call".  ''la'' + ''cmene'' is like a permanent label (Bill is always Bill). ''le'' +
 
''gismu'' is more like a temporary label~--~I have something in mind, and choose to call it "woman".  Probably she really is a woman, but with ''le'' this  doesn't have to be so~--~we could be talking about a transvestite or a stone that looks a bit like a woman.  There are other articles which can show that it's a real woman, or a typical woman or whatever, but we'll leave those alone for the time being.
 
One more word is sometimes necessary when using ''gismu'' as ''sumti'': ''cu''.  This doesn't carry any meaning, but separates the ''selbri'' from whatever comes before it.  It's not necessary with ''cmene'', because they can't run
 
over into anything else, but ''le ninmu klama'' doesn't mean "The woman goes"; ''ninmu'' and ''klama'' get run together, with the result that it means "The woman-type-of goer" (maybe a female traveler).  What we say
 
instead is
 
''le ninmu cu klama''
 
'''IMPORTANT!''' ''cu'' does NOT mean "is" (as in "The woman is going").  In fact it doesn't mean anything~--~it's just there to indicate that there's a ''selbri'' coming. You can also use ''cu'' after a ''cmene'', but it isn't usually necessary.  Similarly, you don't need ''cu'' after ''mi'' (I / me), ''do'' (you, the person I'm talking to) or any words like this ("pro-sumti", in Lojban jargon).
 
==== Exercise 3 ====
 
Add ''cu'' to the following Lojban sentences where necessary, then work out what they mean.
 
# ''la klaudias. dunda le cukta la bil. ''
 
# ''le karci sutra''
 
# ''la kamIL. cukta''
 
# ''mi fanva la kaMIL. la lojban''
 
# ''le prenu sisti''
 
# ''le ninmu cliva''
 
# ''la .istanbul. barda''
 
# ''mi tavla la mari,as. ''
 
# ''la meiris. pritu la meilis. mi''
 
# ''le cipni vofli ''
 
# ''crino''
 
# ''ninmu''
 
=== Changing Places ===
 
We've seen that if we don't need all the places (and we rarely do), then we can miss out the unnecessary ones at the end of the ''bridi''.  We can also miss out the first place if it is obvious (just as in Spanish).  However, it sometimes
 
happens that we want places at the end, but not all the ones in the middle.  There are a number of ways to get round this problem.
 
One way is to fill the unnecessary places with ''zo'e'', which means "something not important".  So ''la suzyn.
 
klama la paris. la berlin. zo'e le karci'' tells us that Susan goes to Paris from Berlin by car, but we're not interested in the route she takes.  In fact ''zo'e'' is always implied, even if we don't say it.  If someone says ''klama'', what
 
they actually mean is  ''zo'e klama zo'e zo'e zo'e zo'e '' but it would be pretty silly to say all that.
 
Most people don't want more than one ''zo'e'' in a sentence (though there's nothing to stop you using as many as you like). A more popular way to play around with places is to use the '''place tags''' ''fa, fe, fi, fo'' and
 
''fu''. These mark a ''sumti'' with a certain place, no matter where it comes in the sentence.  For
 
example,
 
;''la suzyn. klama fu le karce'':Susan goes in the car / Susan goes by car
 
''fu'' marks ''le karce'' as the fifth place (the means of transport).  Without ''fu'', the sentence would mean "Susan goes to the car."
 
With place tags you can also swap places around.  For example,
 
;''fe le cukta cu dunda fi la klaudias.'':The book was given to Claudia.
 
Again, you probably don't want to overdo place tags, or you'll end up counting on your fingers (although they're very popular in Lojban poetry~--~place tags, that is, not fingers).
 
A final way to change places is '''conversion''', which actually swaps them round, but we'll leave that for another lesson.  There are no rules for which method you use, and you can use them in any way you want, so long as the person you're talking to understands.
 
=== Summary ===
 
In this lesson we've covered the following points:
 
* The basic ''bridi'' structure.
* The difference between ''cmene'' and ''gismu'', and the article ''le''.
 
* The place structure of ''gismu''.
* ''cu'' to separate ''selbri'' from ''sumti''.
 
* ''zo'e'' to fill missing ''sumti'' places.
* Changing places with place-tags.
 
Although there is a lot more to Lojban sentences than this, you now have the basics of Lojban grammar~--~the rest is just a matter of adding things on to it~--~different articles, tags, times, numbers and so on.
 
=== Answers to exercises ===
 
==== Exercise 1 ====
 
# ''lojban'' - ''cmene''
 
# ''dunda'' - ''gismu'' (give)
 
# ''.ankaras.'' - ''cmene'' (the capital of Turkey)
 
# ''mi'' - neither, it's a type of ''cmavo'' (structure word) called a "pro-''sumti''", a word that stands in for a ''sumti'', like an English pronoun stands in for a noun
 
# ''cukta'' - ''gismu'' (book)
 
# ''prenu'' - ''gismu'' (person)
 
# ''blanu'' - ''gismu'' (blue)
 
# ''ka'e'' - neither, it's a ''cmavo'' or structure word, meaning "can"
 
# ''dublin.'' - ''cmene'' (the capital of Ireland)
 
# ''selbri'' - neither, it's a ''lujvo'' or compound word
 
==== Exercise 2 ====
 
# ''karce''<br />x{SUB()}1{SUB} is a car/automobile/truck/van [[jbocre: a wheeled motor vehicle|a wheeled motor vehicle]] for carrying x{SUB()}2{SUB}, propelled by x{SUB()}3{SUB}
 
# ''nelci''<br /> x{SUB()}1{SUB} is fond of/likes/has a taste for x{SUB()}2{SUB} (object/state)
 
# ''cmene''<br /> x{SUB()}1{SUB} (quoted word(s] is a/the name/title/tag of x{SUB()}2{SUB} to/used-by namer/name-user x{SUB()}3{SUB} (person)
 
# ''sutra'' x{SUB()}1{SUB} is fast/swift/quick/hastes/rapid at doing/being/bringing about x{SUB()}2{SUB} (event/state)
 
# ''crino''<br /> x{SUB()}1{SUB} is green
 
# ''sisti''<br /> x{SUB()}1{SUB} ceases/stops/halts activity/process/state x{SUB()}2{SUB} [[jbocre: not necessarily completing it|not necessarily completing it]]  
 
# ''cmima''<br /> x{SUB()}1{SUB} is a member/element of set x{SUB()}2{SUB}; x{SUB()}1{SUB} belongs to group x{SUB()}2{SUB}; x{SUB()}1{SUB} is amid/among/amongst group x{SUB()}2{SUB}
 
# ''barda''<br /> x{SUB()}1{SUB} is big/large in property/dimension(s) x{SUB()}2{SUB} as compared with standard/norm x{SUB()}3{SUB}
 
# ''cusku''<br /> x{SUB()}1{SUB} expresses/says x{SUB()}2{SUB} for audience x{SUB()}3{SUB} via expressive medium x{SUB()}4{SUB}
 
# ''tavla'' x{SUB()}1{SUB}<br />talks/speaks to x{SUB()}2{SUB} about subject x{SUB()}3{SUB} in language x{SUB()}4{SUB}
 
Note the different place structures of ''cusku'' and ''tavla''.  With ''cusku'' the emphasis is on communication; what is communicated is more important than who it is communicated to.  Quotes in e-mails frequently start with "do cusku di'e" (''di'e'' means "the following") as the Lojban equivalent of "You wrote" (''ciska'' - "write" - places more emphasis on the physical act of writing).  With ''tavla'' the emphasis is rather more on the social act of talking~--~you can ''tavla'' about nothing in particular.
 
==== Exercise 3 ====
 
# ''la klaudias. dunda le cukta la bil.''<br /> Claudia gives the book(s) to Bill.
 
# ''le karce '''cu''' sutra''<br />The car(s) is/are fast.
 
# ''la kamIL. cukta''<br />"Camille" is a book.
 
# ''mi fanva la kaMIL. la lojban''<br />I translate "Camille" into Lojban.
 
# ''le prenu '''cu''' sisti''<br />The person(s) stop(s) [[jbocre: whatever it was they were doing|whatever it was they were doing]]
 
# ''le ninmu '''cu''' cliva''<br />The woman/women leave(s)
 
# ''la .istanbul. barda''<br />Istanbul is big. (an understatement~--~it has a population of over ten million)
 
# ''mi tavla la mari,a.''<br />I talk to Maria.
 
# ''la meiris. pritu la meilis. mi''<br />Mary is on the right of Mei Li, if you're facing me.
 
# ''le cipni '''cu''' vofli''<br /> The bird(s) flies/fly
 
# ''crino''<br /> It's / they're green.
 
# ''ninmu''<br /> She's a woman / They're women /There's a woman / There are some women In sentences 1, 3, 4, 7, 8 and 9, ''cu'' is possible but not necessary.  In the last two sentences, ''cu'' is impossible, since it has to separate the ''selbri'' from the ''sumti'' that comes before it, and there are no ''sumti'' here.
 
Note that I have translated these sentences in the present tense (since in English you have to choose a tense) but they could be in any tense, so ''le cipni cu vofli'' could also mean "The bird flew", for example.  We'll look at how Lojban expresses tense in later lessons; just remember that you don't actually '''need''' it~--~normally it's obvious whether an action takes place in the past, present or future.

Revision as of 17:07, 4 November 2013

(originally from tu'o and relevant to discussions of tu'o and lo'ei and similar creatures.)

I don't understand how the meaning of the bridi can be determined without quantifying over the underlying set, unless the set is one-membered. Perhaps the idea is simply that the quantifier is left unspecified so that it is glorked from context? That seems reasonable, but it's not equivalent to lo'ei (e.g. mi nitcu tu'o tanxe is not equivalent to mi nitcu lo'ei tanxe). --And Rosta

If the set is one-membered, then you are still quantifying over that one member, though perhaps a bit trivially. I think that the idea (as xorxes has expounded) is that the quantifier is deleted, resulting in the intensional meaning of the selbri being added to the main bridi without extensionally quantifying over the underlying set. This is just what happens with lo'ei
mi nitcu lo'ei tanxe is defined as mi kairnitcu le ka ce'u tanxe. There is no quantification over lo'i tanxe there, only a quantification over lo'i ka ce'u tanxe, which, like every property, whether it obtains or not (or can possibly obtain), is inherently a singleton. -- Adam

If we think of 'sets' as groups instead (collectivities, = Lojban 'masses'), then we can just refer directly to the set/group without quantifiying over its membership. Regarding your analysis of tu'o and lo'ei, I don't get it. lo'ei always seems to reduce to lo with a narrow scope within that of some implicit predicate. In other words, I think I have a rough idea of how lo'ei works, but I can't make sense of your and xorxes's analysis of its working. Maybe this discussion should move to the lo'ei page, if you're arguing that tu'o is equivalent to lo'ei. --And Rosta

Mathematical sets cannot be used directly, and mathematical sets are what lo'i does. Also, in Lojban you cannot just use groups/collectivities/masses directly without quantification. For individual gadri (lo/le/la) you must pick out (quantify over) individuals from that group. For mass gadri (loi/lei/lai) you must pick out (quantify over) parts of that mass. xorxes's rewrite of broda lo'ei brode to kairbroda le ka ce'u du lo brode was an intermediate step, I think, in the ultimate goal of getting to kairbroda le ka ce'u brode. As xorxes says, if you don't understand that, take sisku, which is a primitive in standard Lojban
mi sisku le ka ce'u tanxe (=mi buska lo'ei/tu'o tanxe) doesn't involve any quantification over lo'i tanxe, it merely uses the meaning of tanxe as expressed in the property le ka ce'u tanxe, and adds it to the predicate sisku. --Adam

I don't mean that lo'i broda is a group rather than a mathematical set. I mean that if we conceive of categories as *groups* of individuals, then we can either quantify over the membership, or refer to the group directly. (This is how lei/loi ought to work.) Regarding sisku, as I've been saying in the discussions on Jboske, I don't understand sisku tu'o ka ce'u broda except as a way of expressing troci tu'o du'u co'e lo broda, so for me the analogy with sisku does not help at all in explicating the notion of null quantification. I understand that lo'ei is supposed to "use the meaning of tanxe as expressed in the property le ka ce'u tanxe, and adds it to the predicate sisku", but I can't make sense of this. We haven't found any examples from English that don't reduce to lo (or loi'e). --And Rosta