Hacker News new | past | comments | ask | show | jobs | submit login
Loglan – a predicate logic-based language for humans (wikipedia.org)
98 points by entelechy on Oct 5, 2018 | hide | past | favorite | 37 comments



I think this should mention Lojban. The spin off with a larger community.

In my younger yearsi tried to learn some Lojban. But it was hard. The extreme regularity and brevity of the vocabulary made it difficult to remember words. Also the grammar has a lot of features, special particles allow for flipping word order. Such stuff is actually very confusing.

What i really liked about it is that it allows you to phrase things as precise or as vague as you intend.


It's like the APL of constructed languages.


It's more like Prolog with a maximum arity of 5 and about a thousand particles with complicated effects. The reference text says with some pride that it is "as simple as C++."


That will be Ithkuil with its weird complex rules and alien writing system made from a vast character set.


APL has a very simple ruleset once you sit down and learn it. The character set is large, though (a positive in my opinion, but not for everyone).


Or perl.


  What i really liked about it is that it allows you to phrase things as precise or as vague as you intend.
Is it possible to get some examples of this?


The tense-space system is a good example. By default, a sentence like "mi tavla"/"I talk" doesn't convey any temporal information. But you could add a tense particle like "pu": "mi pu tavla"/"I [past] talked". But that doesn't really tell you how long ago it was, so maybe you want to add "zi" to indicate that it was recently that you spoke: "mi pu zi tavla"/"I [past] [recent] talked" which might be interpreted as "I was just talking". But where were you doing that? Perhaps you were just talking downstairs, so you'd add "ni'a" to indicate that it was below: "mi pu zi ni'a tavla"/"I [past] [recent] [down] talked" -> "I was just talking down there". But maybe you do this all the time? Let's add "ta'e" to indicate habitually. "mi pu zi ni'a ta'e tavla"/"I [past] [recent] [down] [habitually] talked" -> "I used to talk down there all the time". But maybe I want you to know it was kind of a problem or that I did it too much? In that case let's add "za'o" to indicate that the event went on too long: "mi pu zi ni'a ta'e za'o tavla" -> "I used to habitually talk down there too much."

I like this example because: 1) I've only used four or five of the 28 sections of Chapter 10: Imaginary Journeys, which concerns the tense system, 2) this is only the tense system, not any of the copious other systems Lojban has 3) this shows you the kind of particle-soup that I think Lojban can easily devolve into, and 4) eventually, a real Lojbanist will arrive and correct everything I said above, which is an important part of the real flavor of the language. Being less snarky, my suspicion is that there is a real art in choosing which information to add, since none of it is obligatory (unlike English), and it is quite possible to add information in Lojban which would require additional clauses in English, that makes it an interesting mental exercise to interpret.


When I say "I talk to you" in English it's non-optional to have a temporal tense (the above is in the present) but optional to have a spatial tense (you don't know where it is happening).

In lojban you can say "mi do tavla" (or "mi tavla do") and that doesn't specifically give the temporal or spatial tenses. But you can add both to any degree.

So "mi pu tavla do" means I spoke to you in the past, "mi puzu tavla" means it was a long way in the past, and so on.

More details: https://lojban.github.io/cll/10/1/


Actually, "I talk to you" is only habitual; it doesn't have a temporal deixis outside of the hypothetical (although there is a weird narrator's mode in which it could be a sort of present tense).


Would "Me talking to you" be an equivalent of "mi do tavla", without temporal tense? Or did I miss the point?


You missed the point, but it's a hard point.

Suppose I utter {mi do tavla .i ba go'i}. {ba go'i} means "the previous utterance, logically, but with a future tense". English has no satisfactory translation for this entire phrase; the closest we might come is, for example, "I'm talking with you. In the future."


yes, there are some things that you can convey more accurately or precisely in lojban but... i think this is a poor example. Also, I think you're forgetting that translation is about intent not literal word / phrase conversion. The intent of "i'm talking with you." followed by something that indicates that the prior sentence takes place in the future translates very clearly into "i will be talking with you in the future" You can get all picky about verb tenses and say "oh english doesn't have that tense" but a) it's not true ("I am talking with you in the future.") and b) it ignores the point of communication. To communicate concepts between entities. The concept as that in the future communication will happen between you and I. "In the future we are talking." or "We are talking in the future." "talking" isn't temporal it's active. It can be future "we will be talking" or past "we were talking" or present "we are talking".

I'm sure there are some things which can't be translated nearly as clearly into English, but i don't think this is one of those things.


{mlatu} "kitty!"

{oi mabla skami} "Ugh, fucking computers."

{po'o fi'ure lo ro pipno ku zasti} "Only half of all pianos exist."

{cicricfoi bartu} "It's a jungle out there."

{lo se tsarainaitrusi'o ku ta'e na'e pensi} "Fascists typically aren't thinking."

{za'a bridi} "I observe the nature of some logical relation, probably uttered or written somewhere."


I have always held a fondness for {.iu.uinai} ("I am unhappily in love") both for its expressiveness and for the nuance

> {oi mabla skami} "Ugh, fucking computers."

Now I want that on a t-shirt ...


> this should mention Lojban

It mentions it in the first paragraph's last sentence.

> Loglan is the first among, and the main inspiration for, the languages known as logical languages, which also includes Lojban.


I've always had an interest in invented languages and spent some time lurking the conlang/worldlang/auxlang communities. They are a very eccentric and confrontational bunch of people, to put it mildly

The most interesting auxlang project I ever came across was Unish (https://en.wikipedia.org/wiki/Unish), mainly because it is the only one I know of created at a university by trained linguists. It's word selection was highly methodical (http://www.unish.net/introduce/e_unish_Word.jsp) unlike most invented languages which are based on the personal tastes of the inventor.

Check out the book In the Land of Invented Languages by Arika Okrent for a fun and interesting look at the whole topic.


My highschool latin teacher - a highly idiosyncratic individual, not to put too fine a point on it - constructed not one, but two completely artificial languages, although to be fair, one was a derivative of the other, sort of the way many modern tongues are of latin.

The thing is, these were thorougly developed and well thought out constructs - the 'classical' one of them in particular being logically structured and built up from fundamental particles. Felt very much like real languages from what I saw of them. And as probably his only pupil ever to express an interest, believe me, I saw quite a lot. I can still picture him coming to a halt somewhere on the main street, scribbling in his ever present notebook if some new idea had presented itself. And I knew a fresh presentation would be coming up...

Idiosyncratic to begin with, and sliding into decidedly bunkers. He was pensioned off early from his teaching position, and later disappeared on a mountain trek. His languages went with him.


> It's word selection was highly methodical unlike most invented languages which are based on the personal tastes of the inventor.

Another language like this is Interlingua.

https://en.wikipedia.org/wiki/Interlingua


Loglan was for research purposes, IIRC. For actual usage, what little exists, it’s largely supplanted by Lojban.


I think there were license issues involved also. Logan authors wouldn't let the community grow and own the language.


There is also one more interesting specimen - Ithkuil[1].

[1] https://en.wikipedia.org/wiki/Ithkuil


There was a fascinating article posted to HN a little while ago about Ithkuil's creator's relationship with its users.

https://www.newyorker.com/magazine/2012/12/24/utopian-for-be...


The 2015 HN discussion of Loglan [0] included links to XKCD [1] and its debunking [2].

[0] https://news.ycombinator.com/item?id=10075196 [1] https://xkcd.com/191/ [2] https://mw.lojban.org/papri/File:xkcd_comic_debunker.png


This Slate Star Codex post has more substance for its non-existent philosophical language than all of Loglan/lojban. It is a controversial statement but I’ll stand by it:

http://slatestarcodex.com/2013/05/06/raikoth-laws-language-a...

There is a lot of very interesting things that could be done with a human language that had a precise function mapping 1:1 utterances to semantics, which syntactically required specifying source of knowledge and uncertainty, and supported strongly typed noun phrase constructions.

Unfortunately that’s not loglan. The only thing loglan rigorously defines was the syntax and morphology, which ironically is basically a solved problem for natural languages today. Context is all you need to parse a sentence as well as a human, or better, and rigorously defining context was explicitly excluded from the scope of loglan/lojban :(


It's funny how Turkish already is kind of close to a conlang like Lojban


How so?


This might be referring to the regular grammar of Turkish.

In Turkish, if you have a plural noun in the ablative case (roughly meaning 'out from'), you show this by attaching the plural suffix and the ablative suffix to the noun. For example, 'ev' means 'house', '-ler' is the plural suffix, and '-den' is the ablative suffix, so 'evlerden' means 'from (the) houses'.

In Latin (and most Indo-European languages that preserve case), on the other hand, you have one marker for 'ablative plural', that's completely distinct from (most) other case-forms of the plural, and from the ablative singular. So 'domus' means 'house' (in the nominative singular -- there are no zero-marked nouns in Latin, which means that every noun has an explicit case+number marker), 'domibus' means 'from (the) houses', and 'domū' means 'from the/a house'.

This also applies to verbs. In Latin, you have one single marker on the verb that signifies tense, aspect, mood, and person. Passive verbs conjugate completely differently from active verbs, for example. In Turkish, you generally have one marker per thing signified.

(English has lost most of this inflection, but we still have -s for third-person singular present active.)

There's also a lot more irregularity in Turkish than in Latin. Latin has five declensions for nouns and four conjugations for verbs -- so 'domus' takes a completely different set of case markers from 'Rōma'. In fact, 'domus' doesn't really fit into the declension system at all -- it's entirely irregular. (Think English strong verbs.)

The thing about this is that it's actually Latin that's crosslinguistically weird here, rather than Turkish. Most languages are more like either Turkish (lots of affixes, each of which with one and only one meaning) or Chinese (not many affixes at all) than Latin (affixes which mean several different things at once). Most Indo-European languages are to some extent like Latin, but most non-Indo-European languages aren't very much like Latin.

Turkish also doesn't have grammatical gender, or any sort of noun class system at all. (Grammatical gender is one type of noun class system, common in the Indo-European languages.) But noun class systems aren't cross-linguistically uncommon.


Unless I’m missing something, that’s not per se what loglan/lojban sets out to achieve. Does Turkish have precise, unambiguous syntax and morphology? Meaning there is never (other than context dependent semantics) any way for one sentence to be parsed two different ways?


Of course Turkish has its ambiguities and was not designed with the same ideas in mind as Lojban.

But it is in many ways more precise syntactically than other languages and almost completely regular. This is also due to some "cleansing" of the language from Persian and Arabic influences in the early 20th century.

Imho with languages the elephant in the room is context. I don't see how the additional rigour in syntax compared to Turkish would be scoring points for lojban, when the contextual problems that both languages face are not solved.


That was my point above about semantics :)


Semantics are hard. There's a long-running slow-burning debate in the Lojban community about the degree to which BPFK, a guiding body for Lojban, is expected to formalize the semantics and philosophy of the language.

That said, this sounds like a shitpost. Syntax and morphology aren't solved problems for natlangs, and while Lojban's parsing might be messy, it is nonetheless far less ambiguous than English.

Maybe a few Lojban words will enlighten you as to the power of the language?

{ga'i} a marker for formal rank and formal register/diction

{ko'a} a variable

{goi} reassign names and variables

{ce'u} a marker for holes under {ka} abstractions

{ka} definitely not lambda! An abstractor for relations

{ti'e} "I hear that..." a marker for unverifiable hearsay

{da'i} "supposing..." a marker for counterfactuals

Finally, here's something that SSC doesn't have (yet?), a tutorial: https://mw.lojban.org/papri/la_karda


I'm quite familiar with lojban, thank you. And you must have a high bar for content if a critique is considered a "shitpost."

Syntax and morphology are largely solved problems at the frontier of NLP. If you think that this is not the case, then I must ask by what metric do you make that judgement? We have machines that are able to transcribe human speech and infer intent with lower error rates than humans for the same task. QED.

Formal semantics was hard at the time that loglans were codified. It was hard until very recently in fact. But that is no longer the case: Davidsonian event semantics is now expressive enough to give exact, unambiguous semantics for every desired utterance. There are still many unsolved problems in the field of semantics, but they are problems of the sort "what's the best way to represent..." not "how do we reconcile the two main theories giving totally opposite results."

For an overview of the modern state of formal semantics I suggest this work, which successfully dissolved the last lingering major conflicts in the field:

https://link.springer.com/article/10.1007%2Fs10988-014-9162-...


> {ti'e} "I hear that..." a marker for unverifiable hearsay > {da'i} "supposing..." a marker for counterfactuals

Isn't there a language spoken in Africa with a similar concept? I remember vaguely that there was a minor scandal some years ago because a politician said something like "Unfortunately, we couldn't finish the project in time" and the translator used the wrong form, so it sounded like they never intended to do so.


Bulgarian has a verb form for the former.

The latter is commonly encompassed by the subjunctive mood in Romance languages, although that also covers other meanings, depending on the specific language.


first thing that came to my head was Newspeak


Lojban is the inverse of Newspeak.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: