From Lojban
Jump to: navigation, search
veljvo, a tanru turned into a lujvo (via the rules of seljvajvo).

Not every word in the veljvo needs to be represented by a rafsi in the lujvo. A gismu which contributes no places to the place structure of the lujvo, or a cmavo like ke or bo which is there to make the tanru correct, may be omitted if the resulting word isn't likely to be confused with something else. Sometimes even a SE word can be omitted, if the concept is important enough and any other arrangement of the places wouldn't make sense. This occurs, for example, in the word brivla.

The veljvo is written with omitted components in brackets or parentheses. The place structure uses abbreviations such as c1, using the first letter of the gismu the place comes from. If one place in the lujvo comes from multiple gismu, the respective places are separated with =.


This is the form sometimes used when defining a new lujvo:

lujvo: veljvo: brief translation: place structure

An example (from the GNOME translation):

  • nermutmi'i: nenri (ke) mucti minji: applet: mi1=mu1=n1 is an applet for use mi2 running within program n2

Sometimes the ralju brivla deep structure can be provided too.


  • Adam:
    • I question this. Put all your cmavo and gismu in your lujvo, split it into a tanru, or use a more general concept. Whether a word is likely to be confused with something else is hard to tell without context, and may often be a cultural bias.
  • Neither the lujvo list nor what there is so far of the dictionary server has any mechanism for showing omitted words in the veljvo. It would be nice if they did, because trying to make lujvo without allowing for omitted components might cause one to give up on seljvajvo and just fudge the place structure.
    • nitcion:
      • ... or adjust the veljvo accordingly.
        • Which often leads to unwieldy lujvo. It's not like omitting lujvo components is anything new or revolutionary: the Book demonstrates it with sheepdog = lanme [jitro] gerku = lange'u.
          • rab.spir:
            • It is revolutionary. That is called an "anomalous lujvo" in the Book.
              • "Anomalous" does not mean "bad". lange'u is a perfectly reasonable lujvo.
                • Not necessarily bad, but certainly non-seljvajvo.
    • Jay:
      • jbovlaste has a free-form-ish field for the components of a lujvo which is intended to contain a "+" separated list of the components of the lujvo in question. It is separate from the place mapping of the lujvo, which is intended to allow you to produce a list of all lujvo which have mluni2 mapped to their x3 place, for instance. The "components" field of jbovlaste could easily allow for lujvo to have some of their components bracketed to indicate they did not contribute rafsi to the final form. (FWIW, I've spent many a man-day wracking my brain over the design of jbovlaste, just because you can't see it yet doesn't mean I don't support it.