Return-Path: Message-Id: <9208051602.AA28331@relay1.UU.NET> Date: Wed Aug 5 15:49:46 1992 Reply-To: CJ FINE Sender: Lojban list From: CJ FINE Subject: Semntics of abstractors X-To: Lojban list To: John Cowan Status: RO X-From-Space-Date: Wed Aug 5 15:49:46 1992 X-From-Space-Address: cbmvax!uunet!cuvmb.bitnet!LOJBAN Abstractors (NU) have three components to their function: - a syntactic one (they take a sentence and function as a selbri) - an abstract semantic one (they take the meaning of a sentence and deliver that of a selbri) - a specific semantic one (eg 'event', 'experience') For the following discussion, I am going to assume that "su'u" lacks the last of these, and treat it as having only the abstract functions; each of the others is then equivalent to "su'u" plus a specific semantics. In what follows I will render "su'u" as "chunk", to indicate its purely abstract (syntactic and semantic) function. On this basis, I believe that NU can be analysed thus: le nu mi sipna = le su'u mi sipna kei voi fasnu (I use "voi" because there is a clear sense of "which I am choosing to treat as an event for present purposes" - it's not perfect, because "voi" is restrictive as well as subjective - this exposition really needs a subjective incidental) This formulation only works when the abstraction is used as a description. In substantive bridi, it looks something like ta nu mi sipna = ta su'u mi sipna vau fi'o ve skicu lo fasnu "That is-the-chunk: (I sleep), described-as an event" (the fi'o is attached to the su'u-selbri as an extra place) Hereafter I shall use the "voi" form throughout, for convenience. Some of the 10 specific NU are easy to render this way: 1) NU lenu mi sipna = lesu'u mi sipna voi fasnu ("described as an event") 2) LI'I leli'i mi sipna = lesu'u mi sipna voi se lifri ("as something experienced") 3) ZU'O lezu'o mi sipna = lesu'u mi sipna voi se zukte ("an action") 4) SI'O lesi'o mi sipna = lesu'u mi sipna voi sidbo ("an idea") Slightly harder, just because of the difficulty finding a suitable selbri, are: 5) DU'U ledu'u mi sipna = lesu'u mi sipna voi se cusku ("something expressed") or voi se djuno ("something known") or voi fatci ("a fact") 6) MU'E lemu'e mi sipna = lesu'u mi sipna voi mokca ("point in time") or voi mokca fasnu ("point-in-time event") The English rendering of "mu'e" as "achievement" suggests that "se snada" belongs in there, but I think this would be a mistake. 7) ZA'I leza'i mi sipna = lesu'u mi sipna voi ze'e ranji ("that continues over a period") With the last three, I have problems. 8) NI I have had difficulty with this word for a long time, being unsure what it meant, and suspecting that it was extracting something from the bridi, rather than 'chunking' it (hence my answer to Iain that the difference from "jaila'u" was very small). This analysis has just shown me where my problem was: I have been trying to make it leni mi sipna = lesu'u mi sipna voi klani which doesn't make any sense to me, and therefore I was looking for another "klani" somewhere in the chunk. However, if this is rendered leni mi sipna = lesu'u mi sipna voi se klani ("as something with a quantity") it makes much more sense. Thus leni mi sipna cu so'imei = lesu'u mi sipna cu se klani li so'i which is still not specific as to whether it is the frequency, the length, the intensity, the noisiness, or some other aspect of my sleeping which is many, but is in line with the other NU words. 9) KA I have had similar concerns with KA, though slightly less clear-cut. Again, the transformation now seems to me to be leka mi sipna = lesu'u mi sipna voi ckaji ("as a thing characterised") rather than "voi se ckaji" as I thought. 10) JEI "jei" gives a special problem, as it is not the chunk at all, but the result of an operation (truth evaluation) on the chunk. I don't think that "lejei mi sipna" is *any sort of* "lesu'u mi sipna", as I have been using "su'u" here. One might try: lejei mi sipna = leni ledu'u mi sipna kei cu jetnu = lesu'u ledu'u mi sipna kei jetnu kei voi se klani = lesu'u lesu'u mi sipna kei voi fatci cu jetnu kei voi se klani but in fact I believe that "jei" is about THE ONLY THING IN THE LANGUAGE which delivers a truth value, and I don't know how to deal with it. (Logical connectives do so only at a meta-level - the meaning of a connected item is a sumti, a bridi, a jufra etc according to the operands of the connective) This analysis leads me to several conclusions: 1) Despite my earlier reservations, "ka" and "ni" do have sensible, well- defined semantics. I am not certain, however, that they have always been used consistently with these as I have defined them. 2) "jei" is an oddball and needs careful looking at to see whether it even belongs in NU. 3) In general, the semantics of NU is conclusively different from that of jaiBAI (though they may be practically close in many cases) 4) Some of the NU relate to their root brivla via x1 and others via x2 - unfortunate, but probably justified on the basis of both convenience and history. 5) There is no reason to forbid negation and connection of NU, as it can be expressed as negation and connection of the bridi in the VOI. It may not often be useful, but I can think of, for example "za'ijenaimu'i" 6) There is a reasonable interpretation to be given to a generalised NU- creator, ie xu'u selbri xou bridi = su'u bridi [vau] fi'o ve skicu lo selbri (assuming "xu'u selbri [xo'u]" is the generalised equivalent of NU) eg the well known book entitled le xu'u salpydizyfa'a ke nalmatma'e sutryjvi xo'u la .iecu'as. kuctra selcatra co'omi'e kolin