From @YaleVM.YCC.YALE.EDU:LOJBAN@CUVMB.BITNET Wed Oct 6 11:37:13 1993 Received: from ELI.CS.YALE.EDU by NEBULA.SYSTEMSZ.CS.YALE.EDU via SMTP; Wed, 6 Oct 1993 05:39:02 -0400 Received: from YALEVM.YCC.YALE.EDU by eli.CS.YALE.EDU via SMTP; Wed, 6 Oct 1993 05:38:39 -0400 Message-Id: <199310060938.AA00326@eli.CS.YALE.EDU> Received: from CUVMB.CC.COLUMBIA.EDU by YaleVM.YCC.Yale.Edu (IBM VM SMTP V2R2) with BSMTP id 3031; Wed, 06 Oct 93 05:36:52 EDT Received: from CUVMB.COLUMBIA.EDU by CUVMB.CC.COLUMBIA.EDU (Mailer R2.07) with BSMTP id 4630; Wed, 06 Oct 93 05:39:37 EDT Date: Wed, 6 Oct 1993 10:37:13 +0100 Reply-To: Colin Fine Sender: Lojban list From: Colin Fine Subject: Re: Subcategorisation (Was :response to John(lojbab) on sarcu) To: Erik Rauch Status: RO X-Status: Lojbab says ++++++++> Note that Colin often says that Lojban makes the distinction between sets, masses, and individuals mandatory. By the above reasoning this is also not true. "le se cmima" is a set, but doesn't look like one, and "le gunma" is a mass and doesn't look like one. Too much of these distinctions is tied up in the the too-unsystematic, too idiosyncratic, and unbaselined and continuously evolving place structures. Thus I will not push for an explicit or mandatory distinction between du'u and nu levels of abstraction in usage, even if there probably is one. This will undoubtedly lead to some abstraction level confusion in the case of words like sarcu. >+++++++ Our two positions are not inconsistent. I now claim that the distinction between set, masses and individuals is an obligatory grammatical SUBcategory. The level of abstraction is another such (I'm still wondering about this idea of 2nd level abstractions). There are some words and structures which subcategorise accordingly (terbri which must be filled by a set, for example). But not all contexts do subcategorise: the terbri of 'du' for example are obviously not subcategorised. But I must disagree with your examples. "le se cmima" is not a set. It is one or more sets, each set taken as an individual. "le gunma" is not a mass. It is one or more masses, each mass taken as an individual. The pamoi terbri of 'se cmima' (ie the remoi terbri of 'cmima') subcategorises for a set. '(ro) le (su'o)' takes the items subjectively selected by the selgadri one by one: it does not subcategorise, so the items may be sets, masses or individuals. Thus (for example) 'lei se cmima' does make sense, of a sort - it means 'the mass of sets (with members ...)' The subcategorical distinction is lexical and in a sense semantic; but it is not an intrinsic semantics of the thing described, it is in the semantics of how the thing is being related to the rest of the sentence. It is perfectly possible to describe the same collection of (say) people as 'le prenu', 'lei prenu' and 'le'i prenu' - but the claims that can truthfully be made about these descriptions are in general different. All of the above goes for abstraction. The pamoi terbri of 'fasnu' is subcategorised for abstraction, just like the pamoi terbri of 'rinka', and indeed the pamoi terbri of 'nu'. The form is not enough to tell us whether a sumti is +abstract or not: we also need some lexical information (and, with pro-sumti, some anaphor as well). But this does not vitiate the fundamental significance of these subcategories to the language. 'le nu mi limna' is a sumti which is +specific -veridical -mass -set +abstract 'le fasnu' is a sumti which is +specific -veridical -mass -set +abstract but you need the lexical entry for 'fasnu' to know this On the other hand 'le xamgu' is +specific -veridical -mass -set with no specification for 'abstract'. There is a further issue here, on the extent to which abstracts are further subcategorised (eg, actions, states etc.) I haven't really been following the discussion on this one, but I think it centred on the semantics, and hasn't discussed the subcategorisation status. Colin