Lojbanic Jars

In LojbanLanguage, a jar does not exist. Why do we need a separate word to keep jar and bottle separate? Lojban botpi 'bottle' covers all closable containers. But what if I make a jar without a lid? Then it's not closable, so it doesn't botpi (botpi is both a verb and a noun) without a cap. It is a patxu, a container whose caplessness doesn't matter but which is often translated as pot.

Why does English need to differentiate?

If you don't know the difference between a bottle and a jar, I can't help you.

So all botpi patxu, but not all patxu botpi! THERE'S an oversight for you!

I have no idea why you call it an oversight. There are many words with relationships like that among them, in any language.


It should be noted that there are some Lojbanists who see a distinction between botpi and patxu, other than botpi's cap, and find the cap place of botpi clumsy, and therefore propose completely ignoring it.


But English-speakers do recognize the difference between a bottle and a jar. Generally, a bottle has a definable "neck" at the top, and a small opening. A jar, on the other hand, has a wide opening at the top. A bottle is nearly always used to contain liquids, while a jar can contain liquids, or discrete solid objects (like cookies) or granular solids (like sugar). There are differences, so why shouldn't there be two different words? (It is still understood, of course, that bottle and jar both "inherit" from "container".) -- MikeSmith

English language "pots" often do have lids. Bottles and jars are still bottles and jars if their lids are missing (although they might be referred to as pots: a jam-jar minus lid with filled with soil and with a flower plated in it might be referred to as a "pot"), but their lids are ones that attach firmly. Take a beer bottle with a traditional crown cap. Pop the cap off. It's now going to be very difficult to get the cap back on, the bottle no longer closes. But the bottle is still a bottle. Back to our jam jar plant pot. Probably it has a screw cap, so you can see the helical ridges that the cap engages with. These would seem to be tied up with it's "jar-ness", in support of the wide neck. Aren't bottles things that we idiomatically pour out of, while jars we reach into, irrespective of closure? You might pour a thick source out of a wide-necked jar, though. Pots: tea pots, mustard pots, paint pots; all with closures. But closures that have a temporary feel to them, opened and closed often while the pot is in use. Salt pots often have temporary, non-sealing closures that are opened very infrequently, and not at all in normal use. Whereas bottles and jars have connotations of sealing and storage. Pots often are used for the same thing many times, whereas bottles and jars might be reused for something else once their original function is done with. Liquor is stored in bottles, but a fruit jar used to store liquor is still a jar, and still a fruit jar, even if the liquor is poured out of the jar, pouring being an bottle-like activity. Where do we draw the distinction between a jar without it's lid, a pot, a mug, a bowl and a vase. Vases can be quite like bottles. I used to use champagne bottles as vases. See the confusion that the natural language speaker gets into with categories? And see the potential for flexibility and richness of expression they get with it?

The Logban advocate at the top seems to suggest that Lojban completely divides containers into two sets: botpi that close, and patxu that don't. If GeorgeLakoff is right about WomenFireAndDangerousThings this is exactly the way that humans inherently don't think about categories. It's difficult to see how it could be a helpful feature of a language for it to work that way.

Well, I haven't read WomenFireAndDangerousThings, but I assume that it's based on observations of people who speak only natural languages, and maybe based on analysis of natural languages. If LojbanLanguage were to just simply copy the natural language system, then it would end up being just a glorified relexification of EnglishLanguage or StandardAverageEuropean, which wouldn't be worth the trouble to learn. The whole point of LojbanLanguage is that it's different from other languages. LojbanLanguage normally distinguishes different words by their PlaceStructure?, so with-a-cap/without-a-cap is a pretty Lojbanic distinction. (Though I agree that this particular one is probably of little use, you never know how StandardAverageEuropean may be influencing your thoughts.)

It would be interesting to know what a Lojbanistani who's read WFaDT has to say. WFaDT is indeed based on how people use natural language (although most emphatically not just English; the category identified in the title comes from an Australian Aboriginal language). And that's fair enough, I think. It's conclusion is that from the way people use natural languages we can form hypothesis (and verify them) about how people think about the world. Categories are an important way that people think about the world. From the description given here Lojban seems to use what Lakoff calls a "classical" and "objectivist" view of how people form categories. He uses a very powerful argument to show that this theory of categories is the easiest to reason about, but is very little used by people in practice. While it's not difficult to imagine how it might be a benefit for Lojban to be unlike StandardAverage<your language group here>, how is it an advantage for it to be unlike the way that people (irrespective of their native tongue) think about the world?

Well, that is kind of the point of Lojban - people don't think in unambiguous grammars, either.

OK, that's the point, why is it an advantage?

Since we don't really know what's "natural" or "hardwired" or whatever, it's advantageous to see if another system works just as well, or even better. If we base the grammar only on things we find in natural languages, we'll never know whether that's the extent of possible human thought, or if it can be expanded.

Ah, except that the cognitive scientists think that they are exactly on the track of what is natural and hardwired, and as described above, what Lojban leads its speakers to do is neither. Check out The Big Book of Concepts for an up-to-date review of the field.


I've only studied a little lojban, but I'm not sure it's object-oriented. -- NickBensema

Well, in LojbanLanguage you can describe the botpi to indicate whether it has a neck or not, but unless it matters it is not significant, really. If I tell you I have a jar of water, do you really care that it's a jar and not a bottle? If it matters, I'll specify, but LojbanLanguage assumes that it won't matter. If I need to specify whether it has a lid, that's harder in EnglishLanguage but it probably has more of an effect on its use than its shape.
All languages make distinctions that are lost on others. Many European languages distinguish "to live (to inhabit)" from "to live (to exist)": wohnen/leben, habiter/vivre. German has two words for chair (Stuhl/Sessel) [the latter means something like "armchair", I think]. Obviously somebody speaking these languages thinks the distinction is necessary (or did once). If you design a language, you inject your own ideas about what kinds of distinctions are important, and I think you are fooling yourself if you believe your system is inherently more logical than any other. Actually, the logical language refers to its unambiguous grammar.
If I decide to make a container, and I come up with an intermediate shape halfway in between bottle and jar for my design, then when I go make it whether it is to be a bottle or a jar changes its final shape. Have you ever seen a container in between the two? In Lojbanistan you might, because they are both botpi.

But what do you do with a jar whose lid has been misplaced? All distinctions have gray areas, and that doesn't make distinctions flawed.

botpi fo noda or patxu, since a patxu is not inherently capless. Here are the place structures:
Botpi: x1 is a bottle/jar/urn/flask/closable container for x2, made of material x3 with lid x4
Patxu: x1 is a pot/kettle/urn/tub/sink, a deep container for contents x2, of material/properties x3

whoever started this wrote patxa, but the gismu is actually patxu so I have changed that.

So a jar changes its class when it loses its lid, and yet you feel this distinction reflects usage better than that between kettles and tubs? You do some odd things in Lojbanistan. As for the general differences between strong and weak classification schemes, both have different advantages - what one has in comprehensiveness, the other has in distinctiveness - and neither is inherently better, else we should either have an independent word for every vessel ever made or a single noun for all of them. It depends on which is more convenient for the situation. English distinguishes jars and bottles, remember, because over the centuries people have found it useful to distinguish them, and after a short time Lojban speakers might likely invent words to distinguish them too.

The thing that has to scare the Lojbanistani is that, someday, some Lojbanistani will invent a computationally ambiguous language structure, and it will come into common use. Languages don't just evolve new words. They develop new syntaxes.

No, I don't think it scares the average Lojbanist at all. It is explicitly predicted that the language will eventually evolve on its own, and seeing how it evolves, and whether that evolution introduces ambiguities, etc., may help to reveal the nature of language. (Also, people use/misuse LojbanLanguage in ambiguous ways right now, but we'll still learning.)

Wait a minute! How do you say 'urn' in LojbanLanguage?

Does Japanese have a special word for "urn"? But Lojban and Japanese both have methods of either borrowing the word from English, or using native ideas to describe an urn. Lojban has a very powerful method of describing metaphors and modifications (for instance, "a big type-of lidless jar") called tanru, which respect for my audience's interest level forbids me to detail further.

I have an Acme KleinBottle on my desk. Which is it, botpi or patxu? ;) Personally, I go with botpi fo zi'o, but other lojbanists are likely to disagree.

I'll bet they would. Klein bottles are surfaces, not really containers of any sort.


Generics are by definition ambiguous.


The really interesting thing about botpi is that if it doesn't have anything in it, it doesn't botpi [actually ca'a botpi] (or patxu, or vasru or anything else). It can only ka'e botpi, i.e. potentially be a bottle (when something gets put inside it).

In other words, Lojban has words for objects performing an action, but not words for objects capable of or intended to perform an action? That seems like an oversight to me...

No, any word in LojbanLanguage can refer either to performing the action or potentially performing the action. botpi by itself may be a actual bottle or a potential bottle. However, if it's actually a bottle, (which can be made explicit with ca'a botpi), it must contain something, just as it must actually have a lid. But even if it's lost its lid and is currently empty, you can still call it a botpi, presumably meaning ka'e botpi (a potential bottle), since marking tense is optional in LojbanLanguage, and ca'a/ka'e is considered a tense in LojbanLanguage.

The fact that botpi has a place for contents and lid, which are not generally considered necessary to the concept in English is specific to botpi. (Since a bottle containing nothing is still considered a bottle in English, whereas, for example, a father of no one is not generally considered to be a father.) There are many other words whose place structure is more intuitive.

Another possibility is to considered the second place of botpi (the contents) to be referring to potential or intended contents, but that doesn't seem to be how it's normally interpreted.

So, is an actor a [translate actor to Lojban] all the time, or only when he is actually on stage?
 Yes.
There we go, then (unless that's the flippant yes). English provides a way of distinguishing a person who is acting and a person who is an actor. So do the few other languages I'm familiar with (personally I think Latin is especially good with participles). The distinction seems really useful and I'm not sure why Lojban would want to take away the second option like that.

Actor is xeldraci (there are probably other possibilities too). If it's an actual actor and not a potential actor, then there has to be a play, a genre, a playwright, and an audience (don't ask me, I didn't make up the definition of the word).

I would say that he's only a xeldraci when actually on the stage. However, if he's not actually on stage at this moment, then I can't call him a ca ca'o ca'a xeldraci (actually now being an actor), though I can still call him a xeldraci or even a ca'a xeldraci (actually an actor). It could mean, for example, that he's a ta'e xeldraci (habitually an actor).

Lojban doesn't put distinctions such as performing a function vs. having a job into the grammar, and it doesn't put the noun vs. verb distinction into the grammar (and I don't know why it would want to). If you know what you mean, you can make any of those distinctions and more with compound words and tenses, but it's not an inherent part of the grammar.

The distinction between an acting person and an actor isn't a matter of grammar, it's a matter of vocabulary. English has participles (which are adjectives), like acting, plus words for professions, like actor, because it is very convenient to have both. Indeed the second set is the much more common. But Lojban seems to have only provided the participles - xeldraci for acting and nothing for actor. If people were to start speaking the language, I think they would very quickly find that inadequate, and either co-opt xeldraci for the second or invent a new word. One could only hope they come up with a regular system for the second, as English sure didn't.

In English, the word corresponding to "acting" is always a participle (adjective/verb, whatever), and the word corresponding to "actor" is always a noun. That's why I say it is a part of the grammar, though I realize that there aren't any special endings/tenses to indicate the distinction.

As you note, one word in English is far more common than the other, implying that the distinction doesn't really need to be made all that often. In general, it's LojbanLanguage's policy not to make distinctions obligatory. jibri means "job", so a professional actor could be identified as a se jibri le ka xeldraci or (making it a single word) xeldracyseljibri. Likewise, "work" is gunka, so someone who works as an actor (whether or not it's his profession) is a gunka le ka xeldraci or xeldracygu'a. Either of those could work for your "actor" as opposed to "one who is acting", but in most cases it's not necessary to be so specific; the meaning is conveyed just fine.

The impression I get from the above is that Lojban does make the distinction, but only provides words for the half less common in English and related tongues. If xeldraci can be used either way, fine by me, but the above comment about botpi only being such when non-empty is then probably incorrect. If new words have to be coined, I suggest that the board of Lojban or whoever makes decisions comes up with a regular way of doing so! Xeldracyseljibri is somewhat long for the more common of the two...

lojban is baselined, so the board of lojban actively rejects all proposals for change at this point. Also, lojban doesn't use capitals.


A tenuous consensus was once reached, with the understanding that an emptybottle is not a botpi but a botpi kunti ('bottle empty-thing'); & this solution can be generalized to other dilemmas, if one does not want to use the problematical & controversial sumti-deleter zi'o.

EditText of this page (last edited June 25, 2005) or FindPage with title or text search