In a message dated 3/19/2002 9:52:05 AM Central Standard Time, gordon.dyke@bluewin.ch writes:1){roda de zo'u li da su'i de du li no} Apparently only {da, de, di} . {xy} and so on are anaphoric pronouns and will pick up other things. BUT in a clearly MEX environment, they function as variables, for reading formulae. Question: How tell that this is a clearly MEX environment. I think it is, since there is a formula to read: <1)AxEy(x + y = 0)> Further, the parser (do check this always) rejects {li da} out of hand, so use either just {da}, etc. throughout or {li xy} <how long do bound variables last, if propositions are logically connected ?> Disputable, but the reasonable answer is: to the end of the scope of its binding. In non-formulaic Lojban we fudge a bit after that. <2)AxAyAz(x + y = x + z = 0 => y = z) 2){roda rode rodi zo'u du li da su'i de li da su'i di li no .inaja du li de li di} this can also be written 2') {roda su'epada zo'u du li da su'i de li no}, but we are just starting and are not working with languages as powerfull as lojban yet .uinai> Same problem with {li da} as before. I think you need {fa} in front of the right-shifted first arguments: {roda rode rodi zo'u du fa da su'i de da su'i di li no .inaja du fa de di} This still does not work: {du} is a two-place argument, so the three place version does not work; you need {roda rode rodi zo'u du fa da su'i de da su'i di ije du fa da su'i di li no .inaja du fa de di} And finally, the parser does not see {da su'i de} etc. as sumti, but takes the {da} and puzzles about the rest. I wonder if it goes better with {xy}, etc. {su'epa de} and I think the default on {su'e} is {pa}, so just {su'e de}. <3)Ey(x = y * y) 3){da zo'u li xy. du li da pi'i da}> Still doesn't get {pi'i da} nor {li da}. I expect that this is true (mut mut) for 4 as well. And 5. <5)Ax(Ez(x = z**2) => Ey(x = y * 4)) 5){roda zo'u de zo'u li da du li de te'a re .inaja di zo'u li da du li di pi'i vo}> An argument for forethought connectives, though I think this works out right. And you could skip the internal preneces. But, perversely, {vo} needs a {li}. I think part of the problem is the mixture of MEX, which is for reading a formula, with ordinary Lojban, which is about saying what the formula means. They do not mix well. I suspect that all the {su'i} etc. expressions need a descriptor in ordinary Lojban, something like {me'o}: {li xy du me'o da pi'i da} seems to work somewhat better (it still fails, but I am begining to suspect that the parser doesn't do MEX any better than we do). |