From snark!cowan Mon Jun 3 12:11:57 1991 Return-Path: Message-Id: Apparently-From: snark!cowan Subject: Re: cleft place structures To: cbmvax!uunet!math.ucla.edu!jimc Date: Mon, 3 Jun 91 11:40:24 EDT From: John Cowan In-Reply-To: <9105312224.AA24764@luna.math.ucla.edu>; from "math.ucla.edu!jimc" at May 31, 91 3:24 pm X-Mailer: ELM [version 2.2 PL13] Status: RO la djim. kartr. cusku di'e: > In a subordinate clause the modified object is > assumed present in the first place after conversion of the clause. I'm > not sure how this is accomplished, but ke'a (right?) represents the > modified object, suitable for explicit placement in other cases (sorry > -- places). Right. >There are also anaphora fi'e (like pa fi'e) to > represent explicitly the N'th place of the containing phrase. Actually, "fi'e" represents the whole of the containing bridi, so you can access a specific place with "le SE fi'e", where "SE" = null for 1st place, "se" for 2nd place, "te" for 3rd place, etc. This works because "le" selects the front place and the conversions bring a different place to the front. Note the distinction between GOhA selma'o, which replicates the bridi, and the dai-series of KOhA selma'o, which are sumti which refer to the bridi. "di'u" is "the previous utterance", whereas "go'i" is more like "ditto". > In my ideal world filled with replicated sumti, the modified object > anaphor would be provided automatically (if not user-provided) very > much like the anaphora for replicating into or out of abstract sumti. Official doctrine here is that "ke'a" is typically the appropriate filler of an elided x1 place in a relative clause. However, this is not automatic; if "ke'a" appears elsewhere, then "da" may be more like it: lo mlatu poi zbasu ke'a loi slasi the cat such-that [something] makes it [the cat] from plastic probably does not refer to a magical cat that constructs itself. > A small point: just what is the replicated object? Example: > > le mlatu poi nenri le mapku > the cat restrictive inside the hat > > Is the occupant of nenri x1 "the cat", a sumti or... > > mi bajra fi'o nenri kuo le kumfa > I run (tag) inside (tag end) the room > is nenri x1 an abstract sumti something like "lo nu mi bajra"? (And is > this syntax right, particularly is kuo needed? I know about ne'i but I > need to use fi'o.) The first interpretation is very natural, but if > the idea of modal operators is to add nonstandard cases to the brivla > relation, then they should relate the phrase arguments (hat, room) to > the superior bridi or s-bridi, not to its x1 occupant(s). The terminator is not "ku'o" but "fe'u"; this is a reserved terminator for "fi'o", but it is usually needed only when the "fi'o"-tag is being used as a tense adverb, not as a marker for an additional place. You misunderstand the way selma'o BAI tags work. The x1 place of a tag, considered as a selbri, is the sumti which follows, and the x1 place is the only place which exists (all others are elliptical). The true interpretation of mi bajra fi'o nenri le kumfa is I run with-interior: the room which means something like I run with the room inside me. whatever that means; in other words, "le kumfa" is the x1 place of "nenri". You want: mi bajra fi'o se nenri le kumfa I run with-surroundings: the room This can best be seen by looking at the causals: "ri'a" means "with cause", whereas "seri'a" means "with effect". It is an easy mistake to make, and several texts have been published (notably "The Open Window" in JL10) which contained omitted conversions of BAIs. > In the first example the cat-thing is an x1 occupant of mlatu and what > I would like to see is that the bridi "x1 mlatu" is related to the hat. > In other words, "the relation of the sumti referent catting occurs > (restrictive) within the hat". This isn't what we usually think of for > a subordinate clause on a sumti, but if you think about what's really > going on, that's really what it means. Sez you, quoth I. :-) The bracketing of "le mlatu poi nenri le mapke" is not (le (mlatu (poi nenri (le mapke)))) but ((le mlatu) (poi nenri (le mapke))) In other words, the sumti is not that-which-I-describe-as-a-(cat-in-the-hat) but that-which-I-describe-as-a-cat which-is-really-in-the-hat. Remember that "le" is non-veridical, and its scope is short; it is limited to just the following selbri. So the cat in the hat may or may not be a cat, but it must be really in a hat (or something I call a hat, "le mapku"). In any event, the "nenri" part is veridical. -- cowan@snark.thyrsus.com ...!uunet!cbmvax!snark!cowan e'osai ko sarji la lojban