John Cowan writes:
And Rosta wrote: > no'o = mo'e zu'i [typical number] > ???? = mo'e zo'e [unspecified number] > ???? = mo'e zi'o ["null" number, no number] > tu'o = mo'e ???? tu'o is mo'e zi'o, IMHO.
The current mahoste glosses tu'o as effectively equivalent to both mo'e zo'e and mo'e zi'o. This leads to ambiguity. The proposal is to deviate from the official mahoste and construe tu'o as ONLY mo'e zi'o. Quantification with an unspecified number can be done by mo'e zo'e.
no'o and tu'o are merely abbreviations for mo'e zu'i and mo'e zi'o respectively. Any ambiguity about the interpretation of tu'o can always be removed by using mo'e zo'e or mo'e zi'o instead.
From a summarizing message to Jboske:
- Strictly speaking, tu'o is meaningless in itself. It is used when a PA or gadri slot must be filled, but we would rather not fill it.
- As with co'e and zo'e, tu'o needs to be replaced by an explicit value before the truth of the sentence can be evaluated.
- Using tu'o griceanly implicates that choice of explicit replacement for it is immaterial: if the choice mattered, the speaker would have used the desired quantifier.
- The choice of quantifier is immaterial precisely when the extension is a singleton set.
- When 'referring' to singleton categories, it is useful to signal this, because singleton categories are largely impervious to logical scope, and keeping track of issues of logical scope is mentally burdensome. Signalling the singletonhood spares users unnecessary mental effort.
- Reasons for lo pa broda and le pa broda not being satisfactory solutions include: (a) arguably, lo pa makes an unwanted additional truthconditional claim about singletonhood; (b) it is perverse to be obliged to make the effort of using extra words when the object is to save mental effort. tu'o broda implicates lo pa broda.
- It is uncongenial to be forced to make redundant lexical choices that make no difference to the meaning. For singleton categories the choice among tu'o/ro/lo/le is redundant, and rather than having to arbitrarily choose one of them, it would be nice to have a word that neutralizes the choice.
- Addendum: For singleton categories the choice among tu'o/ro/lo/le/loi/lei is redundant.
- I don't see a valid reason to bother using tu'o over lo pa here. Also, I generally see tu'o used with du'u. Is there a semantic difference between that and le du'u? It seems as exactly the singleton category, making gadri redundant. What does lo 2 du'u mean? --la xod
- The reason is given under points (6) & (7) above. You may not think them valid, but those who do think them valid can use tu'o. --And Rosta
- I don't find them valid. For every number except 1 we're supposed to use lo ny., and with 1 we don't announce the number but resort to a completely different digit which only implies 1, and hide the gadri? That is inelegant. I don't understand what truthconditional claim is made by lo pa which isn't by tu'o, or why it would be a problem. Complaints about lexical redundancy would lead one to stick with the standard formula instead of taking advantage of tu'o, which was not invented for this purpose at all. --la xod
- For any number, including 1, lo PA will make a truthconditional claim about the cardinality of lo'i broda. As explained in the summary, sometimes we don't want to make a truthconditional claim that there is only one broda, but if there happens to be only one then it is desirable to have a very simple way to indicate it (for reasons explained in the summary). The complaints about lexical redundancy are complaints about unneutralizable contrasts: Lojban makes such categories as tense and number optional, but it makes categories of specificity and distributivity obligatory, even when they are vacuous.
- What other sort of claim is possible in Lojban besides a truthconditional claim? An attitudinal claim? --la xod
- So-called 'presuppositions', which are propositions that are not part of what the speaker is asserting to be true. By default, Lojban does not use presupposition, but it uses it in at least the following cases. (1) Some UI. (2) Within e-gadri descriptions. (3) Within voi clauses. --And Rosta
- What is wrong with making a truthconditional claim concerning the quantification? That if the number turns out to be other than pa, the entire claim is disqualified? --la xod
- There are various things wrong, such as our wish to be helpful to our interlocutor while at the same time claiming no more than we wish to. But a concrete example is that na ku lo pa broda cu brode does not entail no da cu brode, which is what one would ordinarily wish it to entail. --And Rosta
There is also another use of tu'o which is not inconsistent with the above use to quantify singleton categories. Namely, tu'o deletes the quantification of the description in question, leaving only the meaning of of the description to fill the sumti slot. In other words, broda tu'o brode basically claims that brode, but without actually picking out some actual brode to put in that slot. For example:
- mi kalte tu'o pavyseljirna xirma
- I'm unicorn-hunting. lo in this case would imply that some unicorn actually exists, which is false.
- 'mi nitcu tu'o tanxe
- I need a box/I box-need. (Without claiming that there exists an actual box which I need.)
This latter usage is equally legitimate (I think) but only sorta consistent with the former one. In the former usage, tu'o is used precisely when it doesn't matter what you replace it with (out the the various unmarked alternatives -- le/lei/lo/loi/ro/lo'e/le'e), because they all end up describing the same state of affairs. In the latter use... well, I haven't got my head around it enough to say how it differs, but surely it does...? --And Rosta
- Okay, then in the later usage, the quantification is deleted, and so the meaning of the sumti must be determined without quantifying over the underlying set. It doesn't claim that the underlying set is a singleton; in fact it avoids making any claim whatsoever about the cardinality of the underlying set, so it is consistent with the desire to not have to make claims about the cardinality of inherently singleton sets at all. So it is inconsistent with the above statement "tu'o broda implicates lo pa broda", but it is not inconsistent with the motivation for using it in the first usage. --Adam