BPFK Section: gismu Issues: Difference between revisions

From Lojban
Jump to navigation Jump to search
m (Text replace - "jbocre: p" to "p")
m (Text replace - "jbocre: e" to "e")
Line 178: Line 178:
* There is a general problem with having a full set of words for sensory experience; see [[BPFK Section: sensory gismu]] or http://mail.lojban.org/lists/lojban-list/msg33823.html and http://mail.lojban.org/lists/lojban-list/msg33824.html
* There is a general problem with having a full set of words for sensory experience; see [[BPFK Section: sensory gismu]] or http://mail.lojban.org/lists/lojban-list/msg33823.html and http://mail.lojban.org/lists/lojban-list/msg33824.html


===  [[jbocre: experimental gismu|experimental gismu]] ===
===  [[experimental gismu|experimental gismu]] ===


===  [[jbocre: Types|Types]] ===
===  [[jbocre: Types|Types]] ===

Revision as of 12:22, 23 March 2014

jbovlaste gismu Keywords

Occasionally, it becomes obvious in jbovlaste that a particular keyword is polysemous, or that a particular place deserves a keyword, or that more than one gloss words for a gismu make sense, and so on. These are all gismu changes that do not change the meaning at all, only the English language words associated with the gismu.

Since jbovlaste is intended to be the source for an eventual published dictionary, this sort of change is a bit touchy.

Robin Lee Powell hereby request that the BPFK give to LLG board members, and anyone they might appoint as jbovlaste admins, the power to make such changes without explicit BPFK approval, with the proviso that if anyone doesn't like such a change a vote can be called, and the BPFK's will is the deciding factor in these changes.

This request was approved as of 17 May 2004. It was informally agreed that jbovlaste should be modified to allow listing of changes made under this new rule, if possible, and that such changes should be listed somewhere as they are made. The Approved gismu Alterations page was created for this purpose.

Set Places

I (Robin Powell) am of the opinion that the "set" places in various gismu are un-necessary, and that the set-only restriction should be abolished for all of those places (with the obvious exception of gismu that are specifically about sets). Any specification of a group should be acceptable in these places, as that's all that's wanted. See "kampu", for example (especially since sets don't even have properties!).

Radical gismu Change Proposals

Unapproved

the following minor mistakes were found on the lojban dict server jbo->en (mostly concerning the see also-section): --sarefo

  • sfubu: 2x farlu
  • mutce: 2x traji
  • milxe: 2x traji
  • cliva: 2x litru
  • cumki: see also cumki
  • jdima: 2x pleji
  • jarki: 2x cinla
  • jbari: 2x grute
  • jdini: 2x rupnu
  • tcadu: 2x jarbu
  • jalra: cockroaches are not part of orthoptera; check wikipedia
  • curnu: "almost all multicellular invertebrates are indeed wormlike" - almost all invertebrates are arthropods (mostly arachnids + insects), by number of individuals + species. but i guess you *could* say that many/most(?) animal phyla are worm-like.

The following problems were found in the official gismu list:

  • ciste: The definition does not contain a verb. I believe the words "is a" should be inserted before "system".
  • jegvo: The keyword, "Jehovist," is actually a synonym for "Yahwist" and does not mean "pertaining to Judaism, Christianity, and/or Islam"; the correct term to use here is "Abrahamic."
  • jukni: Crustaceans (which include crabs & lobsters) are *not* arachnids, nor are all non-insect arthropods arachnids. Additionally, the definition needs to be clearer as to whether it refers to all arachnids or just spiders (or all non-insect arthropods, if you change it appropriately); they are not the same thing.
  • jurme: "a bacteria"; "bacteria" is plural, "bacterium" is singular
  • kolme: The definition seems to use "bituminous" as a noun even though every dictionary I have checked says that it is an adjective.
  • konju: The notes seem to imply that "ellipse" and "ellipsoid" are valid translations despite the fact that these are not compatible with the keyword "cone."
  • kurfa: The notes seem to imply that "rectilinear" is a valid translation despite the fact that it does not actually mean "a right-angled shape/form."
  • navni: The list of noble/inert gases omits krypton.
  • skari: It is unclear what sort of thing can fill the x2 (a property, a spectrum of light, etc)
  • tadji: The notes advise to see also "tai", yet "ta'i" is clearly what is meant.
  • xanri: The notes should read "...note that x1 is imaginary...". Also, while not exactly an error, the last statement of the notes seems to be a bit of a non sequitur with the previous statement (Should it say "they thereby do not exist"?).
  • jalra: sanjalra ("spanish roach") should be sagjalra (singing roach).
  • lanbi: x3 is missing.
  • prenu: x2 is missing.
  • bancu: The meaning of the x3 is ill-defined. Is it the thing bound?
  • pensi2/djuno3: that's up with that?
  • catlu/viska, zgana/ganse: It looks like the relationship between catlu and viska is that between "I look at the horizon" and "I see the mountain"; so approx: catlu = troci lo nu viska , zgana = troci lo nu ganse; this needs to be clarified in the definitions.
  • finpe: parenthetical remark "metaphorical extension to sharks, non-fish aquatic vertebrates" implies sharks are not fish, but they are, so that word should be removed.

Other Issues

Issues brought up by members of the community:

Uncategorized Other Issues

  • carna, turning, and rotation. Huge mailing list threads: [1]; camgusmis's solution: [2]
  • detri: The x1 is said to be in format [day,{week},{month},year], using {} for who knows what, and canonically forcing a little-endian date (with week number!). Can we change this to big-endian, and resolve this once and for all?
  • The notes for {djuno} and {krici} don't mention {jinvi} at all, and the definition for {jinvi} doesn't mention "belief" or "believe". There's far too much that makes {jinvi} look like it's only for subjective opinions, and not for objective or testable facts. And using {jinvi} with indirect questions ("I know who killed me") seems off.
  • The notes of rivbi claim that fanta is agentive. Assuming rivbi's notes aren't lying, could the fact that fanta is agentive be added to fanta's definition please?
  • kutyje'u: It should be mentioned that logical tautology is meant, the rhetorical device does not make sense here.
  • linsi: The word “length” in the description is very confusing, because the measure is not meant, but a whole string.
  • rlpowell: WRT panra: I find "differing only in" to be colloquially confusing; "with differentiating property" would make it better for me; I'm not sure I can explain why.
  • selci / ratni ?
  • sevzi: To all appearances, it appears to be about the mind, some sort of secular soul. But apparently lojbab intended it as a translation of the prefix "self-" (please see the link, don't take it from me). But how is that any different from {mintu}? How is {mi lo sevzi be mi cu lumci} (using the intended, not apparent, definition of {sevzi}) different from {mi lo mintu be mi cu lumci}? In any case, the definition (and the cf. list!) needs to be rewritten, because it really seems right now like it's a, well, to quote its definition, ego/id/identity-image.
  • simsa shouldn't have "parallel" in the definition. It should be defined as exactly opposite of panra.
  • xrugau is the only place where I see the proposed definition of xruti by xorxes: x1 returns to state x2 from state x3. I think that should also be here.
  • zdile: having "amusing" as a gloss is misleading; "entertaining", as opposed to "comical", is what is really meant.
  • kribacr and others on IRC are using the term ziltau to refer to what the parsers call a tanru-unit. The idea is that you cannot create tanru out of selbri per se - selbri include things like na broda and ka'e brode. Thus tanru are exclusively created out of ziltau. ziltau lack a 'tense slot' or 'negation slot' - these are things selbri own. The question arises - is this completely accurate? Can we say that na broda isn't really a selbri - but instead that broda is the selbri and na is just a negation 'tag' applied that selbri? Clear definition of what selbri are and if the term ziltau is needed would be useful.

Semantic Inconsistencies

  • Can a person be drani by doing the right thing? Consensus seems to be that no, only an action/event/that sort of thing can be drani, that it's more like {mapti}. But the text is horribly unclear.
    • If it's alright for someone like me to comment here, the gismu list does contain several words which all mean very similar things (drani, mapti, vrude, even xamgu, etc.) --Plastic Raven
  • There have been many threads in the past over an issue regarding the definition of {botpi}, which now seems irrelevant as they had been arguing over an old definition that read something to the affect "x1 bottles x2 with x3". While this is completely irrelevant now, the idea behind it is still valid. Is a bottle still a bottle without contents? I'd say yes. Is a bottle still a bottle without a lid? I'd say no (at that point it's a bottle-shaped cup {lo kabri be tai lo botpi}.). However, we should make a language-wide decision. Does a dog stop being a dog after it dies? Does a bottle stop being a bottle when it's not bottling? Does a programmer cease to be a programmer when ey's not programming? --Lindar
    • If it was up to me, I'd define certain places as being ka'e-able: that is, "this place is normally filled by X [where in the case of botpi X would be "what the bottle currently contains"], but even if there is no current or obvious X, the capability is sufficient for the semantics of this word; zi'o should only be used if the capability has been lost". --camgusmis

Abstraction Inconsistencies

  • Regarding fenki and other such, why are such words requiring {nu}, but xajmi and others not. referenced here.
  • What can the x1 of sidju be? An event or an agent or both or something else?
  • gunma
  • Oren has expressed concern about the laces which require abstraction and would like the definitions to be more specific. Instead of (quality/event/state) it should read (ka/nu/za'i).

Ambiguity

  • What does the x2 of botpi ("container for x2") mean? It is about capability or actual contents or what? -camgusmis
  • What carmi/mutce/tsali/vitci (others?) can apply to: [3]
  • cirko: The glosses, "x1 loses person/thing x2 at/near x3" and "x1 loses property/feature x2 in conditions/situation x3" overload the loss of an object and a feature. Is this a useful compression of gismu space, or does this word have two different meanings?
    • Interesting question, but does it have any pragmatic implications? Anyway, I opine that it has two different meanings. I probably feel this way because I'm so used to Lojban words eschewing this kind of overloading, compared to English. E.g. ponse/ckaji (hey, wait, aren't these the same, modulo tense, to cirko?), nitcu/sarcu, djica/seljaxseldjica.
  • vajni: similarly to cirko, objects and events are compressed in the x2 place. In both cases, this is not ambiguous, as the type of one sumti implies the type of other sumti. Thoughts?
  • ctuca: the x3 place specifies “methods” as included as possible referent set, but also requires it to be a du'u-type abstract. It is unclear how a method can also be a du'u-type abstract, and “methods” should probably be removed. On the other hand, there is no obvious way to talk about instruction in skills, as opposed to ideas.
  • What does dukse mean? Can I say "lo cakla cu dukse lo ka titla"? What about {le girzu cu dukse lo tadni} for "too many students in the group"? I suspect that x1 should be object(s) (mass?) and x2 a property? -camgusmis
  • febvi: The gloss, "x1 boils/evaporates at temperature x2 and pressure x3." is a bit like saying "x1 is a cat/mammal of breed/species x2"." (tswett, #jbopre, 6 May 2011) Boiling is the rapid vaporization of an entire liquid, whereas evaporation is a vaporization that occurs only on the surface of the liquid. Should this gloss be corrected to always refer to vaporization?
    • Consider that the layman (such as the writer of this comment!) might not know such a technical difference between "boil" and "evaporate". To such a layman, they are synonyms, and you know how the gismu list loves all synonyms equally. What remains to be argued is how common are people who know the difference; I can only proffer myself as an example of one who did not, until now, know.
  • The x2 of klani is rather non-specific and some really don't understand what it means. I ( lindar ) think it's a numeric value based on how other measure-words are formed.
  • pritu and zunle have a interesting ambiguity. Is the x3 the orientation of the x1 ("which faces") or a frame of reference form the speaker's point of view? Or is it, as the notes say, "the standard of orientation for x2"? Any of these could be argued, but there are cases where any is true and the others are not. See the thread [4] for more discussion here; there's a real problem, with wording if nothing else.
  • It is unclear if farna is about direction or facing; see the thread [5] for more discussion here. {fa'a} is discussed here.
  • It is *very* unclear what the x3 of tubnu is; see [6]
  • What is going on with the x3 and x4 of xrani? How do you fill those? I certainly don't know. -camgusmis < djancak> hmmm i guess you could damage the property of social well being as opposed to physical well being < ksion> Or self-esteem. Or privacy. Etc.

Place inconsistencies

  • At jbonunsla 2011, we were comparing the sumti for ckana (bed) and stizu (chair). Compared to ckana, stizu seems underspecified. It has only an x1 place. What places do the other furniture gismu have? Did we manage to encode what amounts to a sexual joke in the definition of ckana? Or should all furniture gismu be that generous?
  • The word namcu doesn't include a place for the base, where logically one would assume the x2 would be "in base x2". doesn't namcu1 include the base? (there's a cmavo for it, in mekso) --tsani
    • It says "refers to the value and not the symbol". {li pano} is not a symbol for which you can ask what base it is in, it is just a number, a Platonic ideal if you will.
  • place structure of minli.
  • pluja and sampu are antonyms, but only pluja has the x3 (by standard) place.
  • There are 37 gismu that name specific plants or animals. 33 of these have the definition lo broda cu broda lo jutsi. The 4 exceptions are tirxu, vidru, lanme, and civla. lanme has an x3 of flock. It should either be removed, or flock added to kanba and all of the species of flocking birds. As an example of a collective noun, if it is not removed, perhaps all of the species should be reviewed for whether they "flock."
  • Continuing the above, tirxu has an x3 because it is overloading several feline species in a single gismu, and the x3 allows specificity. Should lujvo be created for each of these species with the coat marking, rather than having an x3 place? Conversely, should jalra, jukni, kumte, mirli, smani, toldi have an x3? They're the other gismu with overloaded (i.e, multi species) animals. (note jalra is mentioned elsewhere in this document)
  • There are a number of gismu for family relations which are not necessarily biological. Of them, dzena, famti, tamne, tunba, bruna and mensi have an x3 place for bond/tie, while rirni, mamta, patfu, bersa, tixnu do not. For example "this is my adopted son" may be expressed by {ti bersa mi lo nu (adopt)}, but then the meaning of the x3 place is implicit, as there's no such place in the definition.

Keywords

  • Cannot find clika when searching for “moss”; “moss” should be the gloss word.
  • The definition for {cinba} useth "busses". Zounds, who knoweth that word in lieu of "kiss"?

Lack of places

  • cisma and cmila need more places, so I can say "what a big smile!". -rlpowell
  • tunlo: x2 is absent. The definition should definitely be x1 gulps/swallows x2. -- mi'e la gleki
  • Shouldn't tunta (definition: x1 (object, usually pointed) pokes/jabs/stabs/prods x2 (experiencer] have an x3 place for the locus on x2 where it gets poked?

Lack of gismu

  • There is a gismu for "liver", but no gismu for "stomach" or "spine"? There is a word for "chest" but no word for "back"? Really? mi'e la .camgusmis.
  • We need gismu for day of the week, day of the month, and month of the year. Addresses in calendar-space, not durations. [7] and [8] for details.

experimental gismu

Types

What type is porsi2: is it a binary predicate, or is it a unary function from predicates to numbers/numberish things (the same type as zmadu3)? In either case, is it ascending or descending? To put it another way, we have 4 cases:

porsi((y1,y2,...),P,A) <-> (for all i,j) i<j <-> P(yi,yj)

porsi((y1,y2,...),P,A) <-> (for all i,j) i>j <-> P(yi,yj)

porsi((y1,y2,...),f,A) <-> (for all i,j) i>j <-> f(yi)>f(yj)

porsi((y1,y2,...),f,A) <-> (for all i,j) i<j <-> f(yi)>f(yj)