From lojban@cuvmb.bitnet Wed Aug 07 21:57:47 1996 Received: from punt4.demon.co.uk by stryx.demon.co.uk with SMTP id AA13738 ; Wed, 07 Aug 96 21:57:45 BST Received: from punt-4.mail.demon.net by mailstore for ia@stryx.demon.co.uk id 839391202:15951:2; Wed, 07 Aug 96 05:13:22 BST Received: from cunyvm.cuny.edu ([128.228.1.2]) by punt-4.mail.demon.net id aa15830; 7 Aug 96 5:12 +0100 Received: from CUVMB.CC.COLUMBIA.EDU by CUNYVM.CUNY.EDU (IBM VM SMTP V2R3) with BSMTP id 8809; Wed, 07 Aug 96 00:12:58 EDT Received: from CUVMB.COLUMBIA.EDU by CUVMB.CC.COLUMBIA.EDU (Mailer R2.07) with BSMTP id 3048; Wed, 07 Aug 96 00:12:28 EDT Date: Wed, 7 Aug 1996 00:11:59 -0400 Reply-To: Logical Language Group Sender: Lojban list From: Logical Language Group Subject: Re: CONLANG: Resolution of compounded words X-To: conlang@diku.dk X-cc: lojban@cuvmb.cc.columbia.edu To: Multiple recipients of list LOJBAN Message-ID: <839391175.15830.0@cunyvm.cuny.edu> Status: R On Conlang: >From: CTGNOLI@IMICILEA.CILEA.IT >Subject: CONLANG: Resolution of compounded words > >John Cowan, while speaking about Loglan and Lojban: >> Actually, there are two subgoals with different purposes: >> to be able to divide utterances into words unambiguously, and to be >> able to divide multi-morpheme words into morphemes unambiguously. > >Please can someone shortly explain how this latter works? Lojban multi-morpheme compounds, called "lujvo" are composed of pieces, called "rafsi" (affixes), and when warranted by the word-formation rules, the letters "y", "r" and/or "n" as "word-glue". The use of the latter are completely prescribed: you always can tell exactly which is required, and you can always tell from the structure of the word whether the occurence of those letters is word-glue. You can also tell by the structure of the word that it is indeed a lujvo and not a name, borrowing, root word, or structure word. Furthermore, for each of the root words in Lojban, there is a small and specific set of rafsi uniquely assigned for use in making the compounds. There is no duplication in the rafsi assignments; this means for example, that the three-letter rafsi "bri" is assigned to "bridi" and to no other root word, and "vla" is assigned to "valsi" and to no other root word. These are not the only rafsi assigned to bridi and valsi. As a result, for example, "brivalsi", "bridyvalsi", and "bridyvla" are alternate formulations of the word "brivla" and are essentially considered alternate spellings/pronunciations of the same word - there is by rule no permitted difference in meaning between "brivla" and the other word-forms. When a Lojbanist sees the lujvo "brivla", s/he knows that a) it is a lujvo, b) that it breaks down into exactly two morphemes "bri" and "vla", and c) that the latter morphemes represent particular root words and thus strongly suggest the nature of the meaning of the compound (though that meaning is not itself determined algorithmically). He knows that any of the word-forms that are made from the two words bridi and valsi are equivalent, and hence can reconstruct the word from its components if necessary (a lot of Lojbanists seem to memorize lujvo by memorizing their components rather than the words themselves, since this helps in recalling the meaning). The result is thus similar to the English compound word "grandfather", which breaks down into "grand" and "father" (though the compound does not necessarily mean that the grandfather is particularly "grand"). But English unlike Lojban does not have unique forms for the components, and thus the word could be erroneously broken into 3 parts "grand-fat-her" which would be incorrect as to morphology, etymology and implication of meaning. lojbab ---- lojbab lojbab@access.digex.net Bob LeChevalier, President, The Logical Language Group, Inc. 2904 Beau Lane, Fairfax VA 22031-1303 USA 703-385-0273 For the artificial language Loglan/Lojban, see powered.cs.yale.edu /pub/lojban or see Lojban WWW Server: href="http://xiron.pc.helsinki.fi/lojban/"