Return-Path: Message-Id: Date: Thu, 23 May 91 04:36 EDT From: lojbab (Bob LeChevalier) To: lojban-list Subject: response to jimc on observatives - exhaustively long Status: RO X-From-Space-Date: Thu May 23 04:37:15 1991 X-From-Space-Address: lojbab jimc writes: >At the L.A. group meeting we discussed "observatives", Initially we had >trouble to analyse the meaning of the bare kunbri "nanmu"; we concluded >that it meant "manliness is happening here", but the distinction between >that and "a man", while obviously real, is hard to explain. That selbri instead of kunbri, now. You are correct that this is an observative. For the benefit of jimc and others who may have old copies of The Loglanist: in TL3, Scott Layson (supported by Chuck Barton) proposed that the bare selbri be interpreted as an observative, after the natural language usage reflected in shouting "Fire" ("fagri") upon sight of smoke (as compared to the then current bare selbri as an imperative). (At that time there was no clear way to declare an observative, though "le nanmu" and "da nanmu" (using current Lojban cmavo) were considered. For newer Lojbanists, please forgive my extensive references to old Institute Loglan in this posting for purpose of explaining to jimc, who is newly learning Lojban after having been quite skilled at old Loglan. By putting it in terms he (and other) old-timers are familiar with and can look up in their old books, the lesson will more readily sink in. Scott and Chuck argued that in most languages, an imperative is in some way inflected, whereas observatives are not. We researched further in designing the Lojban version and found that children first learn to speak essentially in observatives: "mama", "doggy", and occasionally in attitudinally inflected observatives: "milk?" (".au ladru"). My invention of the imperative pro-sumti "ko" solidified the change, and this is now one of the two major identifiable differences between Institute Loglan and Lojban, that are not simply expansions of the language or corrections of hidden syntactic ambiguities. (The other is how we handle and interpret "le nu" abstractions, and that change started as a correction of the Institute Loglan ambiguity concealed in the difference between that version's handling of "lepo" vs. "le po" - now "lenu" vs. "le nu". That difference is also why I habitually write that cmavo pair as "lenu" - our analysis showed that JCB's distinctive interpretation for the two words as separate was doubly flawed, and I and other old-timers occasionally need to remind ourselves that "lenu" is interpreted as the old "lepo" and NOT as "le po". But more on that some other time.) "nanmu" is an observative because the selbri has been atypically brought to the front of the sentence. (Indeed in this case it is the only thing in the sentence, but this is beside the point. "klama le zarci" is also an observative.) This movement, and the explicit elliptical omission of the x1 sumti (the "subject") adds strong emphasis to the selbri as THE critical new information be pointed out in expressing the sentence. Other than this strong emphasis, "nanmu" is treated as any ellipsized sentence is, all unspecified sumti are still actually there, but are unexpressed. They thus have the implicit value of "zo'e" (something I'm not bothering to specify because it isn't important in this pragmatic context.) In the normal observative case, with a physical "subject" ellipsized as for "fagri", a more accurate specification of the sumti would be "(pointing) ta fagri"). But "ta" would cause the speaker to look at the pointer (to see what is pointing where, not look for the fire and run - in hearing the observative "karce" while standing in the street, the distraction of looking at the pointer could be fatal.) There thus is nothing about looking at some "manliness is happening here". That would either be the observative "nu nanmu" or "ka nanmu", which are in turn equivalent to "zo'e [cu] {nu [kei]} "Something is-being-an-event-of something (else)s being-a-man. "zo'e [cu] {ka [kei]} "Something is-being-an-property-of something (else)s being-a-man. where different "zo'e"s can have different values. (I use "manhood" in translation of "nu nanmu" and "manliness" for "ka nanmu", but it is not clear from jimc's example which HE intends.) "A man" would be "pa nanmu", which is a shortened form of "pa da poi [ke'a] nanmu" "One something such-that [it] is a man". This is a sumti, not a complete sentence. We rejected this version of observative, as well as "lo nanmu" and "le nanmu" for three reasons. First, as incomplete sentences, the listener has to wait to be sure that the speaker isn't just hesitating before continuing with a selbri: "pa nanmu ...", "lo nanmu ...", "le nanmu ...". THese are equivalent to the trailing-off incomplete English sentences "One man ...", "A man .../Some men", "The man/men ...". In Lojban such incomplete sentences are defined to be grammatical, and as John Cowan said, are typically used to answer "ma" questions. The latter two translations point out that Lojban descriptors make no singular/plural distinction. The second reason is that these descriptors end up being the first word heard, not the selbri. Shouting "A fire!" has less impact than shouting "Fire!". Finally, the versions with the extra cmavo have just that little bit extra grammar and semantic interpretation implicit in the extra word. In Lojban, obviously, the difference between all these is real and quite significant. >But we came up with a better example: > > carvi It's raining > lo carvi Look, raindrops The first English is a reasonable colloquial translation of its Lojban. More exact is "[Something] rains", or "Rain!" The second Lojban, a sumti, is the incomplete sentence "A raining thing/Some rain ... [is doing something]", whereas the English would be expressed in Lojban as "ko catlu .i carvi [dirgu]" or "ko catlu lenu carvi [dirgu]. (The thing(s) raining need not be drops. We say "lo snime cu carvi" = "The snow rains" and "loi mlatu je gerku cu carvi" = "Cats and dogs rain." is NOT a figure of speech. >Now a grammar question: Translate "The cat on the mat eats the rat" >using a modal phrase. > > le mlatu be vi le matci cu citka le ratcu > >Is "be" used correctly here to link "vi le matci" to the sumti rather >than the selbri? Well, it is legal and grammatical and translates to the English, but it probably is not the best equivalent to the English. For old-timers, "be/bei" is the old Loglan "je/jue", and jim instead wants the 'sumti modifier link' "pe" (formerly part of "ji" along with the current "po'u"): le mlatu pe vi le matci cu citka le ratcu The cat identified-by-association at-the-mat eats the rat. The cat (that is) on the mat eats the rat. More accurate would be: le mlatu poi [ke'a] cpana le matci cu citka le ratcu The cat such-that it is-on the mat eats the rat. The cat that is on the mat eats the rat. while Jim's translates more like the English: "The 'cat at the mat' eats the rat.", where "at the mat" is being treated as part of the selbri-description. "pe" and "poi" attach restrictive relative phrases and relative clauses, respectively, and his original English has an grammatically implicit relative pronoun "that", like my versions. Without any link, the "vi"-tagged sumti modifies the main selbri. Such a sumti could be moved anywhere in the sentence without changing the meaning, so I will do so to make the English equivalent translation more clearly distinct from the other versions. vi le matci [ku] le mlatu cu citka le ratcu At the mat , the cat eats the rat. >Assuming it is (and continuing a previous thread), how >does this differ from > > le bajra be ve lo'e dargu cu citka le cinki > The roadrunner (runner via roads) eats the insect (ve = place tag 4) > >(Please don't flame the literal translation; I needed an exact >parallel.) The official linker is "vo be" (digit 4), but would my usage >actually be rejected by the grammar and would it be incomprehensible to >subsequent steps? No flame - the translation was quite good for what you wanted, if not what you said. I don't quite recognize the "previous thread". Only because you have the wrong word for the place tag will this be ungrammatical. You want "fo" for the x4 place tag. The series is "fa" "fe" "fi" "fo" "fu" for x1, x2, x3, x4, x5, respectively. These are the old "Hixson-Bonewitz tags" "pua","pue","pui","puo", and "puu" announced in an early issue of The Loglanist as a means of reordering the sumti places without gawdawful combinations of conversions (then expressed as monstrosities like "fujunu", for which the current equivalent is the grammatical but probably non-useful "tevese".) "ve" is related to "fo" in that both deal with the x4 place. "ve" is the selbri conversion operator that was "ju" in old Loglan. Grammatically, it would go next to a selbri, hence AFTER and NOT before "lo'e". But that would be useless for your sentence, of course. Your parenthesized translation is the better one. I suspect you want "loi" (old Loglan "lo"), and not "lo'e" (old Loglan "loe" per TL 2 and later): le bajra be fo loi dargu cu citka le cinki The runner via roads eats the insect. and this IS the same in semantical structure to the "be" version of the earlier example. In this case, the English translation more accurately goes with this version, and not with the "pe" version. Indeed "pe" could not be usefully used with tag "fo", because the "fo" is not associated with any particular selbri: le bajra pe fo loi dargu cu citka le cinki The runner associated-with x4-of-something-roads eats the insect. while without the "fo", we get: le bajra pe loi dargu cu citka le cinki The runner associated-with roads eats the insect. The runner of roads eats the insect. ?The runner [that is] of roads eats the insect. The English of the "be" version clearly shows that the roads are the "via route" of running. The "pe" version merely identifies the runner as having something to do with roads. And the fact that the colloquial English sentence does not easily accept expansion with the relative pronoun shows that the "be" version is probably closer to the English "roadrunner". The difference is whether the added sumti is identifying the runner by restricting its domain in some way ("pe"), or whether it is describing the runner more fully as part of the description ("be"). Hopefully this distinction is not too subtle. It says more about the two English examples and their implicit logical structure than it does about Lojban, which merely reflects the two logical structures clearly. For a beginning Lojbanist, I recommend using "be" when attaching a sumti that is semantically part of the place structure (As Nora says, "Use the place structures!"), and "pe" for sumti or "poi" for bridi to attach restrictively identifying information that is NOT part of the place structure. Otherwise you are playing with details of English grammar. (The true beginning Lojbanist would probably just use the tanru "dargu bajra", which most people will not interpret as a road with legs that is moving at a brisk pace or any other implausible interpretation. lojbab