 Shoi ro run bau pendo, i ve shuhu ti dich maskina, mi ba tavla ro do lozhbo bau, i ma ne shertu zhbo sel bau, i kuhi mi drashakahi, ri ve shuhu letivni. Okay, that's enough of that. So what is lozhbon? Let's say constructed language, whose origins go way back to the 1950s, originally designed to test the Sapir-Warf hypothesis that the structure of a language determines, or at least shapes, a native speaker's thought process. So think about Newspeak in George Orwell's 1984, which was supposed to not only simplify the English language, but also make it impossible to think unorthodox political thoughts. So it seems that currently the strong version of the hypothesis that language constrains thought is pretty much discredited, but the weak version that language shapes thought has some evidence. Anyway, apparently the inventor of Loglan, Dr. James Cook Brown, decided to claim copyright on the components of the language and to prevent community modification. Sound familiar? So in 1987, the Logical Language Group created lozhbon, which reinvented the vocabulary. So the features of lozhbon include phonetic spelling, basing the grammar on predicate logic, and making the language have no morphological or syntactic ambiguities. So let's see some examples. Consider I run a mile versus I run a program. The word run has different senses here. The first is physical, while the second is computational. But in lozhbon, a word only has one sense. So here's another example. I saw a man with a telescope. It's ambiguous because we don't know whether I used a telescope to see a man, or the man that I saw physically had with him a telescope. Or how about the chicken is ready to eat? Who exactly is going to do the eating here? In lozhbon, the grammar doesn't allow these types of ambiguities. There are no homonyms in lozhbon where the same spelling and same sounds have different meanings. So in English, the word die is both a noun and a verb, and they mean very different things. There are also no homophones, where a different spelling leads to the same sounds as in cellar, the thing underground, and cellar, a person who sells. And there are no homographs, where the same spelling leads to different sounds like lead, as in the metal, and lead, as in a parade. To be fair, English is a natural language, and has all the evolutionary baggage that goes along with it, while lozhbon is a constructed language. But the rules of unambiguity are baked into lozhbon, so in theory at least, an evolved lozhbon still wouldn't have these problems. Because of its unambiguous parsibility, lozhbon would make a great common language between people and machines. It would remove a whole layer of complexity on identifying parts of speech, disambiguating words, and so on. In Ian M. Banks' culture novels, the language used to speak between people and between sentient machines is moraine. And as far as I can tell, lozhbon is as close to moraine as we're going to get. Now don't conflate the language's unambiguity or its basis in predicate logic with spock-type emotionless logic. The grammar is how you say something, but there are no restrictions on what you can say. In any case, the full reference grammar is here, in this book, which is also freely available on the internet, linked down below. This is the 1997 version, and there have been a few changes since then, but nothing in print and nothing very major. The dictionary is fully online and has never been published, in part because while parts of the vocabulary are stable, words can always be added. So before we start, just a word about the alphabet. Lozhbon uses the Roman alphabet from A to Z, except for H, Q, and W. The vowels are pronounced A, E, E, O, U, and Y is pronounced A. C is always pronounced Sh, J is always pronounced Z, and X is pronounced H. Finally, the apostrophe character is pronounced H. Why not use H for that? I think it has something to do with the history of logline. Anyway, the other two characters used in Lozhbon are the comma, which separates syllables, if they would otherwise run together, and the period, which is a mandatory pause. So let's start, as with many things on the internet, with a cat. The word in Lozhbon for cat is Mlatu. The word Mlatu is, in the grammar, really a relation. The definition of Mlatu is, X1 is a cat, or feline animal, of species, or breed, X2. The X1 and X2 in there are placeholders, or arguments, where you put the things that are in the relationship. So if I wanted to say that I am a cat, I would be in the X1 position. In fact, a sentence or statement in Lozhbon is simply one relation, all the arguments that go with it, and any additional tenses that apply to the relation. An example of a tense would be the time tense, so I could say I am a cat, or I was a cat, I will be a cat, yesterday I was a cat, and so on. Another example would be emotional tenses, so I could say happily that I am a cat, or confusedly that I am a cat. There is also a system for modifying relations, so that I could say I am a black cat. There is a system for modifying arguments, so that I could say five of you are cats, and for skipping arguments, so that I could simply say cat. Note that I never have to include any trailing arguments like the species argument. There is also a system for adding arguments to a relation, as in I am a cat sent by the authorities in order to arrest you for crimes against the internet, and many more systems for basically anything you would want. So let's start constructing some sentences. Now before I continue, let me just say that Lozhbon has specific words for grammatical categories. The word for sentence is bridi, and the definition of bridi is X1 is a bridi with the relation X2 and arguments X3. The word for relation is selbri, which of course is itself a relation. X1 is a selbri with relation X2 in language X3. And the word for argument is sumti. So in I am a cat, the relation is a cat is the selbri. The argument I is the first sumti, and the entire sentence is the bridi. The ordering of a bridi is essentially one sumti after another in place order, with the selbri somewhere in there after the first sumti. Normally, you put the sumti first, then the selbri, then the rest of the sumti, but you can move the selbri further down, rather like instead of saying I am a cat, saying a cat am I. In that sentence, there is more of an emphasis on the fact that I'm a cat rather than something else. Now the word for I in Lozhbon is not a relation word, it's a prosumti, meaning that it stands in for a sumti. And the word for I is me. And so we can put together our first full bridi, me-mla-tu. Okay, let's modify the selbri. The word for black, as in the color, is hekri, meaning X1 is black or extremely dark colored. So you can guess what me-hekri-mla-tu means. Hekri-mla-tu is the selbri here. Historically speaking, it means X1 is a black color type of cat, and the specific meaning is left up to the people communicating. Now I never said that Lozhbon was semantically unambiguous. You can always use circumlocution to make the meaning more clear, if necessary, as in I am a cat which is black colored. What if I want to use that X2 position in me-la-tu, as in I am a cat of breed Persian. I am a Persian cat. The word for Persian, as in culture, is kul-nir-farsi. X1 reflects Persian or Farsi culture in aspect X2. There's also a word for Persian, as in the language, ban-fuh-hasu. X1 is the language with ISO 639-3 code FAS, or Persian. Or Iranian Persian as a language, which is ban-fuh-hasu. X1 is the language with ISO 639-3 code PES, or Iranian Persian. But none of these really means Persian as a cat breed, so let's just use it as a name. In Lozhbon we can lozhbonize a name according to certain morphological rules. The rules convert Persian into Persian. Those dots signify mandatory pauses, so you can't run the sounds of the word into other words, otherwise the parser will break. And what I have now is a grammatical class called shmevla, meaning X1 is a name word, meaning X2 in language X3. Now a shmevla is not a sumti, so we can't just use it as X2. We have to convert it into a sumti. And there is a word in Lozhbon which does that, la. It is not a relation word, but something more like an English article, and it turns any following shmevla into a sumti. So our sentence is now mimlatu, la, perjan. Now suppose that someone adds a word to the dictionary for the Persian breed of an animal. Suppose it's juter-perja, meaning X1 is a Persian breed of animal X2. Now I can't simply use juter-perja in the X2 position of mimlatu, because juter-perja is not a sumti. We need a converter word, and there are several. We've seen one, la. One of them is lo, which is like a, or sum. There's also le, which is like the. Lo is nonspecific, that is, you don't have any specific instance or instances of the thing in mind, while le is specific. We usually say my cat is a Persian, instead of my cat is the Persian, so I think lo is the right one to use here. So now we could say mimlatu, lo, juter-perja. But what would happen if I left out lo? I would end up with mimlatu, juter-perja. Here there is no X2 sumti because the sentence doesn't parse like that. Mlatu, juter-perja, would be the salbury, and it would basically mean cat-persion breed. Like there are different types of Persian breed, and one is a cat. Are there any other animals that have Persian breeds? I don't know. Okay, let's try to translate the first sentence of the hobbit. In a hole in the ground, there lived a hobbit. So first, let's tackle the salbury. We're looking for a word live, as in dwell in, not remain alive. That word is khabju, X1 dwells in location or home X2. Now, clearly in this relation, X1 would be a hobbit. We can lojbanize this to a name like hobbit. But someone's already gone and added a word in the dictionary. Reader hobi, X1 is a hobbit of mythology X2. And we've seen that lo is a nonspecific converter of a salbury to a sumti. So now we have lo reader hobi shu khabju. Shu is a word which separates the first sumti from the salbury. It's necessary here for reasons we'll see later. Now we can put a hole in the ground as X2 of khabju. Kevna is the relation for a hole. X1 is a hole in X2. And ground, as in land surface, is tumsve. X1 is the ground of land X2 at location X3. Now we can't just say kevna lo tumsve and then try to use lo to convert that to a sumti because lo converts a salbury to a sumti. And kevna lo tumsve is not a salbury. You can imagine that if we just say lo kevna lo tumsve, then any reasonable parser would just imagine we have two sumti one after the other. We can instead use tumsve kevna, which is a salbury, meaning X1 is a ground type of hole. And it seems pretty clear to me what that means. So we can now say lo reader hobi shu khabju lo tumsve kevna. It just means a hobbit dwells in a hole in the ground. There are two problems with this translation. One is that we're relating a story and that usually means everything is in the narrative past or story time tense. To put a breedy in the past tense, you just prefix it with pu. Pu lo reader hobi shu khabju lo tumsve kevna. A hobbit dwelled in a hole in the ground. A pending key to a tense makes it sticky so that until further notice, you don't have to use the tense in every following breedy. So pu ki lo reader hobi shu khabju lo tumsve kevna. A hobbit dwelled in a hole in the ground. The second problem is that this is very dry. It has nothing of the nuance of the English version which puts a hobbit at the end of a sentence. This is a kind of punch or emphasis to the reader because up until then the sentence was quite ordinary. We could try shifting the salbury. Pu ki lo reader hobi lo tumsve kevna shu khabju. A hobbit in a hole in the ground dwelled. A bit more poetic sounding but to me it doesn't really capture the impact of the original. Now in Lojban there's a way to exchange two places in a relation. The word se swaps the x1 and x2 position in a relation. So if khabju means x1 dwells in location x2, then se khabju means x1 is a location dwelled in by x2. It's sort of equivalent to the passive voice but the English passive voice is more limited compared to swapping x1 with any other argument in a relation. So let's try it. Pu ki lo tumsve kevna shu se khabju lo reader hobi. A hole in the ground was dwelled in by a hobbit. I don't like it. But we can use the rich tense system of Lojban. Lojban has a tense for location, nehi, which means within or inside of. Just because khabju has a place for the home of the dweller doesn't mean we have to use it. So instead we can say pu ki nehi lo tumsve kevna lo reader hobi shu khabju. In a hole in the ground a hobbit dwelled. I like that. Ship it. So what have we learned so far? Bredi, selbri, sumti, a whole bunch of other relation words. Lo, le, and la, shu, se, pu, and ki, and nehi. Now you might have noticed that the relation words all seem to have at least five letters while the smaller functional words have less. This is a result of the word form rules of Lojban. Relation words that have five letters in the pattern ccvcv or cvccv are called gismu or root words. X1 is a root word expressing relation X2 among argument rolls X3 with affixes X4. As of 1997 there were about 1,350 of them. Since then about 185 have been added as experimental. Words not in the list can be formed by combining words that are in the list. The first method is what we saw before. Modify one gismu by another gismu like hekrimlatu is a black cat. Or suppose we wanted a word for android. Maybe renma minji or human type of machine could be used. This grouping is called a tanru. X1 is a binary metaphor formed with X2 modifying X3 meaning X4 in usage X5. Tanru are always ambiguous with the meaning up to context and what the speaker and listener understand. Many times the meaning is fairly obvious as with hekrimlatu. And this is why shu was necessary in lorider hobby shuhabju because otherwise you'd end up with just a sumti, lorider hobby habju. A hobbit type dwell action. Maybe a hobbit lifestyle? Tanru can also be formed from a gismu and another tanru. Consider small black cat. The gismu for small is shmalu. We could try shmalu hekrimlatu but unfortunately lojban groups to the left. So the tanru here is formed of shmalu hekri modifying mlatu. And I have no idea what a small type of black coloring is. Maybe slightly black? Gray? In any case, lojban has a word where you can explicitly group the parts in a tanru. Bo. Bo binds its neighbors into a group. So shmalu hekri bomlatu forces small to modify black cat. Another way we can say small black cat is by saying that the cat is small and black. In lojban the word zhe is used to bind its neighbors into an and group. So shmalu shekri mlatu means small and black cat. That is the cat is small and black simultaneously. What about black and white cat? We can't use zhe because that would mean the cat is both a black cat and a white cat simultaneously. What we mean by black and white is that the cat is black in parts and white in parts. Lojban has the word zhoi which is the mixed together meaning of and. So hekri zhoi labna mlatu would be a tuxedo cat. The next way to form a word or concept is to sort of smash the gizmu together. Gizmu have combining forms and there are a bunch of rules for combining them. Going back to our human type of machine example, renma has two combining forms rem and reha while minji has mihi. Under the rules we can't combine rem with mihi because that leads to a double consonant which is disallowed. We are allowed to insert a y leading to rema mihi or we could try reha mihi. This also is disallowed because all relations must have a cluster of consonants in the first five letters not including y or apostrophe. We are allowed to insert an r leading to reha mihi. We can also drop the final vowel in a gizmu and use that as the combining form. So combining reman with mihi to form reman mihi. The combining forms of gizmu are called rafsi and the combinations are called lujvo. So now we have several types of relations gizmu or root words tanru or metaphorical groupings and lujvo or combined words. By the way the dictionary entry for android is formed from remna simsa which means similar to and minji. Remna and simsa combine to form remsmi which means x1 is humanoid and that combines with minji to form remsmi mihi a humanoid type of machine which is probably more precise than my attempt. The final way to form a word is to borrow it from the native language. This borrowed word form is called a fuhivla. To do this you lujbanize the word then add a prefix formed from a rafsi of the gizmu that describes the general category of word. So if we really hated remsmi mihi we could first use the rules to lujbanize android to maybe chandroy and then combine with a rafsi from machine minji mihi chandroy maybe. So now we have gizmu lujvo fuhivla and tanru. A tanru is just a string of words where the words are gizmu lujvo or fuhivla. Those three types of relation words are called brivla so a tanru is a string of brivla and a celbri in a bridi is a tanru or a brivla. Okay so how about we try translating the first line in a Stephen King novel like the stand which is on everybody's mind. Uh wait that's just sally that's in the preface does does that count? Well okay the context is getting someone's attention. Lujban has a whole bunch of vocatives which are words used to address someone for various purposes. They precede eishmevla. The lujban word for getting someone's attention is juhi so juhi salis. To end this video let's translate the first sentence in Stephen King's dark tower the gunslinger which goes the man in black fled across the desert and the gunslinger followed. The first part seems pretty easy the celbri is going to be something like x1 fleas but there's no dictionary entry for flea. Flea basically means to run away from something. There are many entries including run and the one most interesting is bajliha. x1 runs away from x2 via route x3 on surface x4 and don't forget to start story time. You can run the small words together because they still parse the same. How about the man in black? We have nanmu which means x1 is a man. In black presumably means wearing black clothes. We need a relative phrase such as who wore black clothes. In asumti we can start a restrictive relative phrase using poi. The phrase is restrictive because it narrows down the sumti. Poi is followed by an entire breedie and is terminated if needed by kuho. And finally keha is used inside that breedie as a prosumti to refer to the sumti that the phrase is attached to. So let's put together wearing black clothes. Dasni means x1 wears x2 as a garment of x3. It might be sufficient to just use hekri as x2 just black stuff as no specific garment. So dasni lo tuho hekri. Tuho is a quantifier like a number but it means there is no count not even one just stuff. So now we have puki lonanmu poi keha dasni lo tuho hekri shubajliha. Paho is a directional tense capturing a cross and shantuha means x1 is a sand desert. So puki lonanmu poi keha dasni lo tuho hekri shubajliha paho lo shantuha. You might ask why lo, uh and not le, the. Well we haven't introduced the man or the desert so we have no prior reference to them. Once we've mentioned them we can start using le because then it will be clear that we're referring to the previous mention. Let's deal with the gunslinger followed. Follow in this sense is jersey x1 pursues x2. We should really fill that x2 place with something referring to lonanmu so that there's no doubt about what the gunslinger is following. To refer to x1 of the previous bridi we can use the sumti legohi so jersey legohi. There is no word gunslinger in the dictionary but hey let's use a tanru to combine gun shmashel chahi with man. Lo shmashel chahi nonmu shu jersey legohi. Now to combine the two bridi we use ije. i begins a new bridi. Remember zh from small and black cat. It's the same one but when used after i it connects the two bridi in a logical and connection. So finally we have puki lonanmu poi keha dasni lotuho hekri shu bajliha paho lo shantuha ije lo shmashel chahi nonmu shu jersey legohi. Thanks for watching. Kiheiro Kinzga.