From @YaleVM.YCC.YALE.EDU:LOJBAN@CUVMB.BITNET Fri Mar 26 12:01:46 1993 Received: from YALEVM.YCC.YALE.EDU by MINERVA.CIS.YALE.EDU via SMTP; Fri, 26 Mar 1993 07:51:57 -0500 Received: from CUVMB.CC.COLUMBIA.EDU by YaleVM.YCC.Yale.Edu (IBM VM SMTP V2R2) with BSMTP id 4587; Fri, 26 Mar 93 07:50:41 EST Received: from CUVMB.BITNET by CUVMB.CC.COLUMBIA.EDU (Mailer R2.07) with BSMTP id 5675; Fri, 26 Mar 93 07:51:54 EST Date: Fri, 26 Mar 1993 12:01:46 GMT Reply-To: C.J.Fine@BRADFORD.AC.UK Sender: Lojban list From: C.J.Fine@BRADFORD.AC.UK Subject: Re: TECH: grammar updates To: Erik Rauch Message-ID: ki'ecai djan > > Well, I have now decided to withdraw my support for the "*mo'u" construction. > The formal change is being kept in until a few off-list people (Nora and > possibly pc) have a chance to review it, but I personally have accepted > Colin's arguments, and added one of my own. > > This confusion has deep historical roots, resulting from the fusion of > three or four different Old Loglan word-groups into our uniform BAI. > Historically, only the causals were subject to SE conversion, making them > into "effectuals" (I don't know the proper word for "seki'u" and friends). > The Lojban designers generalized, allowing every BAI to take SE, and allowing > (in principle) any selbri to be used as a BAI-type tag. But the essence > of a modal tag is that it governs only one sumti; it is not a relation > between two or more sumti. .ie There are several unfortunate consequences of that generalisation, as well as the fortunate ones. As I understand it, the SE that precedes BAI is not the same as other SE anyway. As you say, BAI are 1-place, and the relationship between BAI and SE BAI is: if do'e represents fi'o broda then se do'e represents fi'o se broda te do'e represents fi'o te broda etc So the SE is a selector of the appropriate tergismu from the underlying selbri, rather than a converter at all. transforms into something like: > We still have the (pure) forethought connectives like "semaugi...gi" and the > mixed logical/modal afterthought connectives like ".esemaubo", and these > should suffice for the cases, if any there be, where such constructs must > exist. It will always be rather indeterminate, however, what is to go into > the second position of such connectives, because of the inherent one-place > nature of BAIs. > I think it is a mistake to think of these as connectives. I think it is more productive to think of them as connectives plus tcita on the second connectand. This is easy to see for sentence connectives: .ijeseri'abo co'e = .ije (bo) seri'a co'e just as .ijebabo co'e = .ije(bo) ba co'e so I suggest .ijesemaubo co'e = .ije(bo) semau co'e (whatever that means) and xy. .esemaubo .y'y = nu'i xy. nu'u .e semau .y'y Thus the BAI remains 1-place. > I have added fuller support for non-logical connectives: they may now be > used in termsets on a (grammatical) parity with logical connectives, and > they may be used in short-scope (JOIK-BO) and long-scope (JOIK-KE) uses. > These new constructions are allowed within sumti, selbri, and MEX. .i'e > The Nick/Lojbab experimental cmavo "xo'e", which eradicates a place > (so that "da klama xo'e xo'e de di" means the same as "da litru de di"), > has been assigned the cmavo "ne'e" and placed in selma'o KOhA. I'm dubious that this is either necessary (has anybody ever used it) or a good idea, but I don't think it does any harm. > Nick's declefter is now also part of the grammar: it does not have its own > cmavo, but is signalled by "jai" without a following tense or modal, thus > keeping the close link between "jai" and "fai". Its effect is to take the > abstraction normally falling into the x1 place of the selbri and move it > to the extra "fai" place; the new x1 place is one of the places of the > subordinate bridi within the abstraction. (If x1 is not an abstraction, > "jai" is ill-formed). Exactly which place is chosen is unspecified, so > "jai" may be equivalent to "jai gau" or to something else. Example: > > 5) le nu mi catra la djim. cu jensa la djein. > The event-of my killing Jim shocks Jane. > > becomes: > > 6) mi jai jensa la djein. fai le nu [mi] catra la djim. > I shock Jane by the event-of [my] killing Jim. This is very clever (perhaps too clever). For the first time I sort of understand it. I have looked at Nick's suggestion several times, and somehow it has never made its way through to the necessary part of my brain to be understood. I look forward to seeing it in use. > "ma'o" now accepts a full mekso (terminator "te'u") rather than > just a lerfu string. This allows operand-to-operator coercion for > lambda calculus and other unusual mekso. .i'ecu'i > > Lastly, the recurring desire to add new kinds of abstractors to NU has been > achieved without any new grammatical machinery. "su'u", the vague abstractor, > now has the place structure: > > x1 is an abstract nature of of type x2 > > giving us the title: > > 7) le su'u la .iecuas. kuctra selcatra > be lo salpydizyfa'a ke nalmatma'e sutryterjvi > the abstract-nature-of (Jesus is-an-intersect-shape type-of-killed-one ) > of-type a slope-low-direction type-of non-motor-vehicle > speed-competition > The Crucifixion of Jesus Considered As A Downhill Bicycle Race > > as an example. This place structure makes explicit the notion that > every abstractor is a specialized kind of "su'u"; for example, "nu...kei" > means the same as "su'u...kei be lo fasnu", the abstract nature of > (some bridi) of-type an event. .u'e.i'esai Very neat, and a beautiful bit of consolidation. mi'e kolin