From LOJBAN%CUVMB.BITNET@uga.cc.uga.edu Wed May 24 12:41:55 1995 Date: Wed, 24 May 1995 17:13:07 +0300 From: Veijo Vilva Subject: Re: quantifiers on sumti - late response To: Bob LeChevalier Message-ID: la lojbab cusku di'e [...] > There are some limits to what you can put before the bare selbri in a > simple description. We have expanded this to allow for preposed > relative clauses after the Finnish model, and this took considerable > work and debate. A recent proposal to allow preposed "be/bei" > constructs was embedded in last week's discussion - it might work, and > might be useful, to people with a preposed grammar native structure. > Veijo??? But no guarantees we could make it grammatical, and it is not > importnat enough to justify a change if it causes anything more than a > trivial expansion rule (if even then). I haven't had much time lately for even reading the incoming mail but the above caught my eye. I didn't read the proposal to allow preposing "be/bei" but the idea feels quite natural. I think the following additions might work: sumti_tail_A_112 : selbri_130 | selbri_130 relative_clauses_121 + /* preposed linkargs */ + | linkargs_160 selbri_130 + | linkargs_160 selbri_130 relative_clauses_121 /* explicit inner quantifier */ | quantifier_300 selbri_130 /* quantifier both internal to a description, and external to a sumti thereby made specific */ | quantifier_300 selbri_130 relative_clauses_121 + /* explicit inner quantifier and preposed linkargs */ + | quantifier_300 linkargs_160 selbri_130 + | quantifier_300 linkargs_160 selbri_130 relative_clauses_121 | quantifier_300 sumti_90 ; This would make possible sumti like, e.g. (1) le be do prami (Fin) sinua rakastava (2) le re be do prami (Fin) kaksi sinua rakastavaa (3) le poi mi viska ku'o re be do prami (Fin) nakemani kaksi sinua rakastavaa I like the structure (of course :-), and I think it wouldn't mess up anything. co'o mi'e veion --------------------------------- .i mi du la'o sy. Veijo Vilva sy. --------------------------------- >From LOJBAN%CUVMB.BITNET@uga.cc.uga.edu Tue May 23 >From lojbab Continuing to respond to Dylan: >Well, yes, but because of the implicit existential quantification of >{da}, a {noi} clause ends up providing information crucial to >restricting the quantification and understanding the sentence; e.g., {da >noi nanmu cu te cimei} == "There are at least three men." No - that is *exactly* what "poi" would do. You have "Something (BTW a man/men) is a trio". > > Well, there is the classic example {le nanmu cu ninmu}, then it > > would be true that {le nanmu na nanmu}, which means that > > {le nanmu cu nanmu} is false. We could argue ad nauseum about > > ... > >Let me just make one point that may be new, without trying to start a >new discussion: gender and (physical) sex are hardly as dualistic as >most people think they are (and as our culture constantly pushes on us). >Consider transvestites, transsexuals, and hermaphrodites. Because of >those last, in particular, one might reasonably say {da cu nanmu je >ninmu je nimnau}. A transvestite is (normally) either a man or a woman, and hence is an excellent case where one might say "le nanmu cu ninmu" or vice versa. A hermaphrodite is a nanmu joi ninmu and probably a "je" would not be as effective since I think most hermaphrodites do not fully function as either of the sexes. A transexual is of a specific gender by genitalia and a different gender by genetics, so the distinction is one of "by standard", but whichever the standard, the transsexual is one or the other not both. > > In the case of klupe vs. dinko, it is always the case that a screw has thread s, > > annd it need not be the case that a dinko has a point per se (if it does, the > > use of jesni might be in order to talk about it). > >(which also doesn't have a place for the point...) Depends upon what aspect you are talking about - if you mean the TIP, you wany jipno. If you want the fact that the nail/fastener is pointed (i.e. needlelike) you use jesni. But I think the definition is such that a distinct single point is not the essential aspect of dinko. For one thing it might be pointed at both ends. And, in any event, the nature of the point does not determine whether or how well it fastens (it may affect how well it penetrates, but that isn't stated in the definition of a dinko). The nature of the screw threads and of the shaft are both relevant to how a screw works, and how well it works. > > There is also a little history in these object gismu. At one point very > > early on, we identified a clear semantic distinction, a hierarchy as it were > > from tutci-cabra-minji-zukte. Some of the place additions came from a > > mistaken attempt to say that some gismu were necessarily subclasses of one > > of these - e.g. a mruli is always a tutci etc. and it was for brevity that we > > trasfered some places of tutci to mruli. > >Not sure what you mean. Has {tutci} lost places since this time? All >the places of {tutci} seem to be covered in {mruli}. ({to'o} What's >the fourth terbridi of {mruli}? The handle or an agent?) Here are edited gismu entries for the foursome, and mruli (Hey, what about my other gripes (e.g., the motion selbri)? Should I >just ignore places I don't like in text I write?) No. You don't ignore them. There must BE a value, even if it is not convenient or useful or important to say what it is. Thus in an abstract sense, if there isn't a value for from or a to, it isn't farlu. It is merely a motion propelled by gravitational force. lojbab >From LOJBAN%CUVMB.BITNET@uga.cc.uga.edu Tue May 23 >From lojbab I'm running behind (as usual) and muchh of this may have been answered. Dylan: >This sounds good, although I'm still not sure {lo} must necessarily be >verdical. In some cases information must be inferred in the use of >{lo}, as in your use of {lo cnino} I pointed out (unless that was a >misuse); why not the meaning of the description? lo must necessarily be veridical, by definition, as I explained at length. Call it dogma. The recent discussion seems to be that when you use lo in a way that seems non-veridical - you are merely creating a restricted universe of discourse in which it is veridical. Thus "lo {unicorn}" is veridical and invokes a universe of discourse where unicorns exist. I presume that a statement about "lo ci nanmu" refers to a universe of discourse which has an unspecified restriction from the real universe - a restriction such that there are only 3 things that really are men in it. But "lo ci nanmu" DOES refer to men and "lo ci nanmu cu ninmu" (or more properly "lo ci nanmu cu na nanmu") is AND MUST BE false. Whereas "le ci nanmu cu ninmu" or "le ci nanmu cu na nanmu" need not be false. > > > I'm not sure we should be so hasty. {lo broda} can be glossed {da poi > > > broda}. Can {le broda} similarly be glossed {ko'a poi broda} (or > > > perhaps {by poi broda})? > > > > Rather {by voi broda}, or {ro da voi broda}. > >{voi} must have been added since the "Places" paper was written. Is >there more to its use than just the one-sentence description in the >cmavo list? > >In any case, {ro da voi broda} seems odd. There's explicitly a >quantifier there. Does the {voi} cancel it in some way? voi was added solely to allow for non-veridical restriction: "da voi nanmu". There is an x that I describe to be a man. It was added for logical completeness and not because it is especially useful. The fact that you can use a "voi" clause attached to a "lo" or a "le" description is simply another side effect of including its parallelism to poi/noi in other grammatical contexts. It is possible that they may mean nothing, or more likely that what they mean is something no one would want to say because there are so many easier and clearer ways to say it. Since it was added as an extra cmavo in an existing selma'o, not much thought was given to nuance of semantics - only actual changes to the YACC grammar were subject to extremely careful analysis prior to the cmavo list baselining. (Now of course everything is trial-baselined for the dictionary and we are trying mightily to make only changes that are really useful, and then only if they offer no impact to the much-lagging dictionary work. Every one of these ever-lagging 'critical issues' is one more distraction (read 'excuse') to further delay the dictionary). >> It is in 99% of the cases. But you are still allowed to use >> {le broda} to refer to somthing that is not even remotely a broda, >> as long as context makes it clear what you're talking about. > >What's an example of when you'd want to do this? That is, can you think >of a context in which {le broda cu broda} isn't true in any sense >whatsoever? That would seem extremely strange to me. This is a classic of the Loglan project. le nanmu cu ninmu That man over there is really a woman (That man whom we have been describing turns out to really be a woman). (i.e. you have been presuming that she was a man perhaps because her clothes and hair make her appear masculine, but I know better). >Hmm, I need to ponder this more. There seem to be two uses of {poi} >involved: subselection and definition (as with {da}). > > > >OK, I think I understand now. {poi} means subselection, Yes - always. It can be definitional, but the way it defines (as in "da poi" is by restriction (the preferred term to 'subselection') of the universe of discourse to that subset that meets the selected place of the restrictive clause. > while {noi}, >by providing more information, means definition (in an appropriate >context, e.g., {da noi li'o} if {da} is going to be used again). But >if the variable {da} is not used again, the two are equivalent. No. "noi" always *merely* provides additional information. The information is incidental and could always be left out of the sentence entirely without changing the meaning. Instead it could be stated as a separate sentence on the same logical level. e.g. da noi broda cu nanmu = da cu nanmu .ije da cu broda, with the added metalinguistic status that the broda claim is subordinate and incidental to the main line of thought. The comparables in American English (British can be different): John, the teacher, is sitting next to John, the bartender. (poi - here the clauses are distinguishing which John is which) John, who BTW I saw at the store yesterday, is sitting over there. (noi - you presumably know which John I am talking about, and the info about my seeing him is irrelevant to identifying him or to the claim of his sitting - thus the latter is logically equivalent to "John is sitting over there, and I saw John at the store yesterday." >So the equivalent of {le broda goi ko'a} might be {ko'a noi le broda}, >at least if {ko'a} is not already in use. There is no intended equivalence between goi and noi/poi. "goi" is a metalinguistic device to define one undefined sumti (presumably shorter) as being metalinguistically a reference to the other (presumably longer or more complicated grammatically). "cei" does the same for two selbri. One could erroneously say that "goi" means "poi du", but really it means "sei du". We have also said that usage is flexible between having the definition on the left or on the right depending on stylistics. >So what does {voi} mean? What's an example of its use? Per above, voi is a way to accomplish the intensional aspects of "le" i.e. not making a claim about ke'a (in da voi ke'a broda), but merely referencing a description of convenience. Most usages of "lo broda" which are not carefully restricted should properly be "da voi broda" and not "da poi broda" which implies that "ke'a broda" is a complete and true statement of all info needed to restrict the universe of discourse to the intended meaning of "da". If I say "mi nitcu lo tanxe" I am implying that any box will suffice, from a ring-box to a refrigerator-box. If I say "mi nitcu le tanxe" I am implying a specific in-mind box (which may not truthfully fit the predicate ke'a tanxe). If I say "mi nitcu da voi tanxe" I am getting something half-way in between - I don't think it is necessarily a specific box, but the restriction is certainly specific and in-mind and not necessarily veridical. (Does this solve that bloody "any" problem?) lojbab