Date: Sun, 12 Apr 1998 04:21:42 -0400 (EDT) Message-Id: <199804120821.EAA09776@locke.ccil.org> Reply-To: Logical Language Group Sender: Lojban list From: Logical Language Group Subject: Re: Lojban ML: Syllogism and sophism X-To: jorge@INTERMEDIA.COM.AR X-cc: lojban@cuvmb.cc.columbia.edu To: John Cowan X-UIDL: 7307a891bf8edee3f69473d6c97b823b X-Mozilla-Status: 8011 Status: RO Content-Length: 7868 Lines: 168 >>I'm not sure that it is, IF you require no use of human intelligence >>or experience to disambiguate the polysemy. Every word or sentence in >>English ADDs to the ambiguities to be resolved. It is only through common >>sense recognition of the overlaps of related semantic spaces that we know >>which meaning of each word applies. I suspect that this is one of the >>hangups in AI. > >You could prove your assertion very easily by giving an example >of something that can be said in Lojban and cannot be translated to >English to the same degree of unambiguity. Well first of all, since I am still thinking in English it is hard for me to "think of something in Lojban" that does not already have an in-mind English translation. But I dare say that we have some good examples of English ambiguity right here in this discussion. The Lojban would likely be less ambiguous (if done correctly). To wit: > >But I don't see what the difference is between a number on >--More-- >>a scale and a measurement of a property using the units of said scale. > >I don't either. > >>5.3 on the scale of grams IS a measurement of the property/quantity of >>mass. > >Right. Now, is {le ni} the measurement, or is it the property being >measured? Is it {le se merli} or {le te merli}? According to example >5.3 of chapter 11 it is {le te merli}. According to example 5.5 it is >{le se merli}. na'i The x1 of ni is a quantified property measured on scale x2 (of ni broda) The quantified property is not a pure number - but the combination of a pure nusorry - the combiantion of a quantified property and a scale on which it is measured implies a pure number as the measurement. Thus I see no problem with convertion a ni abstraction to a grammatical quantifier using mo'e you are converting a combination of two places. Now te merli is a pure number - one would normally put in that place a quantifier like "li pa" which in itself is rather meaningless. It gains some meaning when combined with ve merli - the scale - to be a scaled number (dimensioned number). Thus, omne could say that 5.3 in chap 11 (p 261 for those trying to find out what we are talking about) is grammatically acceptable but semantically would earn a flunk in a science class. The amount of blueness in the picture must be expressed as a scaled number if it is to be represented as a quantifier at all. You cannot subtract a scalled number from a dimensionless number any more than you can subtract 1 gram from 3 kilometers (different dimensions). So other than as an example of the grammar, I can assign no meaning to 5.3 that is acceptable to science. But the reason it is unacceptable is of the realm of mathematics, and not of language. 5.3 does indeed express "1-B" as Cowan says it does; we just do not know what that means mathematically. Now let me note that the difference between merli and klani is the presence of an agent in merli. Because an agent/observer can sometime affect the value (te merli) being measured, your question is also na'i for that reason - te merli is specific to the measurer for a given quantifiable proerty being measure on a particular scale. It is also a pure dimensionless number - you need le ve merli or le te klani to get the dimension into the picture. So your question more appropriatley might have been whether the x1 of ni broda is le klani or le se klani (the x2 of ni broda is clearly le te klani). I believe that, basing the definition of ni on klani, we want the x1 of ni broda to be le klani, RECOGNIZING that given a measureable property and the scale (and fixing all other variables), the pure quantifier that is the measurement of that quantifiable property on that scale is deterministic. I have no problem with accessing that quantifier, so long as it comes with the dimension, using mo'e. So mo'e leni broda du mo'e leni brode can work even though broda and brode are different things,if the ellipsized scale of measurement on them is compatible (it need not be identical - 100o meters can equal 1 kilometer). Now turning to 5.5, the example is cvlearly set up to parallel with 5.4 for contratoive purposes. But measurement is not always applied to a single place within a bridi - you sometimes choose to measure the relationship itself - so I am not sure that ce'u is always needed or meaningful. In this example, where blanu has but a single place, I see no difference between 5.5 le pixra cu cenba leni ce'u blanu and the phrasing I would use, which is le ni le pixra cu blanu cu cenba There are no examples in the book (that I know of) that deal with more complex selbri than blanu, but I similarly would say le ni ko'a sutra lenu broda cu cenba again not needing a ce'u. I choose to summarize the abstractors in terms of their associated gismu. The Book says that these associations are more memory hook that inferential, but I think that is largely because no one has thought out the inferences that might be drawn and so we do not want to commit to the possible inferneces in advance of said thinking. ni le klani ka le se ckaji nu le fasnu li'i le se lifri si'o le sidbo Now the possibility that one could infer from the Book examples that le ni broda was le se klani is itself due to English ambiguity. I know of no English way to express the meaning which categorically eliminates le se klani while keeping le klani. So this might be the example you asked for at the start of this message. If this example isn't good enough, I refer back to our discussion of djuno, wherein I am not sure that everyone figured out what everyone else was saying in English, in spite of tomes worth of text, but we did significantly better when we stuck to Lojban words and pseudo-Lojbanic English (even if there we still failed to agree). >>> >As I have repeatedly said, my concept of slabu as age i slabu be loi >>>>jmive prenu. >>> >>>But "well known to some alive people" is not equivalent to old as age. >> >>"Known to the living" is how I would prefer to express it. And a person's >>age is indeed how long s/he has been known to the living. > >So someone who has lived for very long but is not known to anyone >living is not old? Every person become known to"the living" at the time of birth or almost immediat immediately thereafter, or they do not live very long. In theory I guess a mother who died in childbirth with no others around might cause the newbor n chiuld to not be known to the living in a technical sense immediately on birth, but if someone doesn;t come along to take care of the child, the child will never grow old. It is part of the nature of being familiar to t "the living" that we are constantly in this state from birth until death (at lea st). Theonly question is whether a fetus is similarly characterized a s "familiar to "the living" because of association with a presumably living preg nant mother. >That's not what I understand by "old". And how come >--More-- >{ni} would take the meaning of "how long" here rather than "how much"? >Is {le ni marji} how long something has been material The key property of slabu is supposed to be length of association/familiarity. Remember that the e-translation ARE English, and there is usu ally some tie to English usage that led to the wording. In the case of slabu, we might use the phrases "old news" or "old hat" for something that is slabu-familiar. We would not use slabu for something that I learned just today, no matter how well I learned it - it isn't THAT sense of English "familiar/well-known", that would permit a short term knowledge to be so described. So the essence of slabu is a length of time. But to make this clear, one could insist that one say leni slabu kei le si'o nanca as distinct from leni slabu kei le si'o grake in case someone wishes to try to measure a quantity of slabu-ness as a kind of mass. lojbab