rant: against sumti qualifiers

From Lojban
Jump to: navigation, search

Rant against sumti qualifiers


Bare NAhEbo is a glorky version of PA LE PA NAhEbo, where NAhEbo is logically a selbri -- "is other than", "is opposite to", &c. Yet to express this underlying meaning we have to resort to {du PA LE PA NAhEbo}. Instead, NAhEbo + sumti should be a selbri. That way we can say "x is other than y" as {x na'ebo y" instead of {x du pa lo ro na'ebo y}. {PA LE PA NAhEbo} remains unchanged except that it cannot be glorkily abbreviated to NAhEbo.


  • There is no meaning for lu'o, lu'i and vu'i among current proposals and CLL that is both coherent and useful. These cmavo are just a waste.
  • lu'a, meaning "is a member of" ({lu'a X} = {PA LE PA lu'a X}), is useful, but, like NAhEbo, should be a selbri. Examples of {me} in CLL suggest that {me} must mean "is a member of", so lu'a is redundant.
  • lu'e, "is a reference to" ("is a text about"?), should also be a selbri.
  • la'e should also be a selbri, but its meaning should be clarified and the candidate meanings should each be assigned to a separate cmavo. If X is an utterance, then "la'e X" could mean:
    • (2a) the full proposition that the speaker intends to communicate, including the illocutionary force;
    • (2b) the full proposition that the speaker intends to communicate, excluding the illocutionary force;
  • This leaves tu'a, which I think is always short for {le ro tu'a}. As a glorkless abbreviation, it can be left unchanged.

Changes to implement the above

  • NAhEbo + sumti becomes a selbri, not a sumti.
  • lu'o, lu'i and vu'i, and lu'a, are scrapped due to different sorts of uselessness.
  • lu'e, la'e and three new variants of la'e move to ME. Concommitantly, the ME+MOI construction is scrapped and the rules for MAI, MOI and ROI adjusted to allow them to be preceded by any operand.
  • tu'a remains the only member of LAhE.

What you want for NAhEbo + sumti seems to be already achievable as NAhEme + sumti, at least for some interpretation of me. --xorxes

  • Yes, so long as me + sumti counts as a selbri -- which is not unreasonable. --And.


The NAhEBO examples look to be archetypical malglico -- they may work, but they are inelegant to the max. NAhE is predicate negation. It assumes that a predicate holds a place in a symmetical structure, such that there is an opposite to that predicate and a neutral position in the structure as well. Of course, this whole works only for a very limited array of predicates (what is the opposite of "chartreuse"? "mauve"? What is the neutral color: colorless? white? black? What is neutral or opposite to "cow"?) Only {na'e} itself seems to have general untility -- meaning "not this property but some other of the same kind." Of course, what kind is meant is left open in many cases: is "cow" about quadripeds, mammals, vertebrates, animals, living things or what? Glorking, I suppose, is called for. The set is also surprisingly incomplete: one expects, since it is scalar, to have at least markers for near misses and far ones, and -- most importantly -- cases where the kind of property doesn't apply at all (color to atoms and spirits, say -- something like {na'i} but with the grammar of NAhE). For theological discussions, it would also be handy to have something for "farther away from neutral" along with "on the other side of neutral."

  • The NAhE series can be supplemented with the CAI intensifiers: je'acai, je'asai, je'aru'e, to'eru'e, to'esai, to'ecai, etc. Maybe {je'ana'i} could work for "does not apply"? --xorxes
    • Well, {je'a na'i} parses correctly, so maybe that will do -- though I can't see what {je'a} does here except introduce a paradox between exactly and not at all. The same applies the strong and weak forms for near misses and far: they are not weak or strong hits, but not hits at all -- but not opposite or neutral either.
      • {je'aru'e} and {je'asai} are weak and strong hits, of course. {na'eru'e} and {na'esai} would be the near and far misses. In {je'ana'i}, {je'a} is within the scope of {na'i} so I don't see a paradox. {je'a broda} is the affirmative of a scale where broda is the positive, and {na'i} says that the scale does not apply. {na'ena'i} would probably work just as well. --xorxes
        • Well, even just {na'i} parses in these positions, but the bracketting suggests groupings that are semantically wrong. It is hard to work out what the scope of {na'i} -- or almost anything else that has a scope, for that matter -- is. I assume it is either sentential or the next chunk (whatever that is, again) that follows. I notice that the obvious ways, at least, do not allow including a sumti in the other-than-ness or inappropriateness modification: We can't readily say "John is other-than-going-to-the store," meaning he is coming from the the garage or sitting on the porch or ...; we can only deny a particular relation to an object: "John is other-than-going-to the store."
          • I take the scope of UIs to be always the immediately preceding word. That is what the formal grammar suggests. (Anything that falls within the scope of that word will automatically fall within the scope of the UI, so if that word heads some longer construction, that whole construction will be covered by the UI. In particular, a UI after {i} will have scope over the whole bridi headed by that {i}.) To include the sumti in the other-than-ness we can say: {la djan na'e ke klama be le zarci}. It seems that the {ke} is needed, as {na'e} binds closer than {be}, but with {ke} the sumti is definitely under {na'e}.
            • No obvious way, as I said -- we can only do it with a defective tanru (i.e., one lacking the modified). As for the scope issue, I know what the parser says, but that is notoriously uninformative of what is going on (see other discussions of the logic of the situation). In some cases it is simply wrong, in others inadequate -- and maybe occasionally right. YACC is for acceptors, not for serious grammar work: it tells you that something is grammatical, but not what its grammar "really" is.
              • Do you object to UI's scope being the previous word (and thus any construction that word may head)?
                • Depends on the UI: most of the UI1s I can make any sense of seem to be sentence length wherever they occur, some of the UI3s make sense with some words or phrases but with others seem to go to sentence, UI2s seem to be mainly sentence, and so on (admitting that what some of these guys, especially some of the UI1s, mean is fairly uncertain). In any case, the bracketting in the parse is totally uninformative, since it is the same for all.
                  • I agree that some of the UIs are somewhat nebulous. My position is that if it makes sense to interpret the scope of a UI as the preceding word, then the scope is the preceding word. YACC does not always give the real parse, but it is not completely arbitrary either, so it is not a bad policy to follow it as much as possible.
                    • I guess my policy is that, since it is often arbitrary, it should never be used as more than a hint, to be overridden by almost any consideration. As for taking the scope of UI to be the previous word, I just find that almost impossible to interpret in most cases -- how can I be hopeful, for example, about a single word (or, rather, its referent)?
  • For example:

a'o la djan pu dunda lo xrula la meris

I hope John gave flowers to Mary.

la djan a'o pu dunda lo xrula la meris

I hope it was John who gave flowers to Mary.

la djan pu dunda lo xrula a'o la meris

I hope it was flowers John gave to Mary.

la djan pu dunda lo xrula la meris a'o

I hope it was Mary John gave flowers to.

    • But the hope is still sentential length; the positioning merely emphasizes some aspect of that hope as being central. That is, it contrasts this hope with a number of other potential ones, varying the indicated place (well, except of course, that it does not say that this is actually my hope). This has at least a pragmatic aspect, maybe even a semantic one (the contrast above), but it does not affect what the scope of the operator is, so far as I can see. This may be just a terminological dispute, since I think we agree on what happens here. I would describe it in terms of the application of a lambda'd proposition to a term (possibly a predicate, of course) -- but all in the scope of a "hopefully" a modifier to the highest order performative.
      • Perhaps then we can say that UIs have sentential scope but also establish a focus on the preceding word/construction. More examples:

la djan la'a pu dunda lo xrula la meris

It was probably John who gave flowers to Mary

la djan pu dunda lo xrula ju'a la meris

It was flowers John gave to Mary

la djan pu dunda lo xrula la meris xu

Was it Mary John gave flowers to?

la djan e'a dunda lo xrula la meris

Let it be John who gives flowers to Mary

la djan dunda lo xrula ei la meris

It should be flowers John gives to Mary

la djan pu dunda lo xrula la meris ja'o

So it was Mary John gave flowers to.

        • Well, the grammarian side of me would say that the position of the UI represents or displays the focus established in the sentence, rather than that it established it, but yes, that is what happens -- just as we always knew it did. Terminology, as I said. pc
          • Lots of things in Lojban we always know and keep forgetting at the same time. {je'ana'i} shouldn't present a problem then: presupposition failure, with focus on the selection of a point on the scale. (I still don't have a clear picture of the scope of UIs in embedded/subordinate clauses.)
            • Too painfully true! I am not sure what "focus on the selection of a point on the scale" means. If {je'ana'i} can occur just anywhere, then I can imagine that where it occurs may say something about where the presupposition failure is: is it, in the classic case, with "stop" or "beating" or "wife"? But I don't see that as fitting into a notion of scale. As for the embedded UIs, it seems to depend on the UI again -- and on the nature of the embedding (abstraction seems different from simple subordination and these from coordinate conjunctive cases). Not that I have any particular ideas about how any of these work, just that it seems unlikely that they all work the same way always.
              • {je'a} contrasts with {na'e} (in vs. out of a category) and also with {no'e}-{to'e} (positive vs. neutral vs. negative on some scale), so I would take {je'a na'i} to say that the presupposed category or scale is not appropriate. The UIs deserve a scope of UIs ew page, I would say.
                • Given that interpretation, that looks right. The same effect could apparently be achieved using {na'i} alone in various places.

What happens when this gets applied to sumti rather than selbri? There are two obvious possibilities: {na'ebo lo broda} = {lo na'e broda}, or {na'ebo lo broda} means that some other sumti referring to things of the same general kind as brodas applies. these are the general and limited versions of the same point and are, presumably, both correct. In the case of gadri other than o's, probably only the second applies, {le na'e broda} is conceptually ill-formed, since the speaker has, presumably, not picked out a bunch of things that he is describing as non-brodas. Similarly {la na'e brod}, even if it were syntactically possible, it would be conceptually hopeless. I suppose that {na'ebo la brod} simply means that some other name is called for -- or at least another description, while {na'ebo le} calls for another description, though not necessarily a {le}.

  • I think {na'ebo la brod} calls for someone other than Brod, not for another name. {na'ebo zo brod} would call for another name. Similarly {na'ebo le broda} calls for something other than the thing I have in mind, not for another description (of the same thing??). --xorxes
    • Sorry for being too brief; it calls of course for someone or something else, under a different sumti (the temptation to merge the two meanings of {sumti} is endemic in Lojban and I fall into it from time to time).

The LAhE rants suggest ignoring what CLL says: that most of these are shortened forms of descriptions pretty much as given. Of course, since we are weaving in and out among at least three and probably five different readings of some of these, it is a little hard to say for sure. If there is a reading in which the complex translations really do hold, I would take that as an argument against that particular reading.