[298] in tlhIngan-Hol
Re: Causatives
daemon@ATHENA.MIT.EDU (daemon@ATHENA.MIT.EDU)
Mon May 4 16:27:15 1992
Errors-To: tlhIngan-Hol-request@village.boston.ma.us
Reply-To: "Klingon Language List" <tlhIngan-Hol@village.boston.ma.us>
From: Ken_Beesley.PARC@xerox.com
To: "Klingon Language List" <tlhIngan-Hol@village.boston.ma.us>
Cc: Ken_Beesley.PARC@xerox.com
Date: Mon, 4 May 1992 12:16:00 PDT
In-Reply-To: "tlhIngan-Hol-request@village.boston.ma:us:Xerox's message of 4 Ma
Mark,
I ran the "QaghHommeyHeylIjmo'" example through my program and got the
following two solutions:
Qagh-Hom-mey-Hey-lIj-mo'
[ mistake/error {NS1} diminutive {NS2} plural {NS3} apparent
{NS4} your(sing) {NS5} due_to/because_of]
Qagh|Hom-mey-Hey-lIj-mo'
[ mistake/error | bone {NS2} plural {NS3} apparent {NS4} your(sing)
{NS5} due_to/because_of]
In the second reading, which is obviously not the intended one, the first two
morphemes are interpreted as two nouns compounded together: "Because of your
apparent mistake-bones." What that might mean is up for grabs, but it is
well-formed. I have found many such ambiguities while running examples from
the book. Take a simple "yaSpu'", and you get two solutions:
yaS-pu' [ officer {NS2Sent} plural[sentient] ]
yaS|pu' [ officer | phaser ]
Even worse, boQpu'
boQ|pu' [ aid/assistance | phaser ]
boQ-pu' [ aide {NS2Sent} plural[sentient] ]
boQ|pu' [ aide | phaser ]
+boQ^pu' [ they [-obj?] {V} assist {VS7} perfective ]
-boQ^pu' [ she/he/it-him/her/it/them {V} assist {VS7} perfective ]
*boQ^pu' [ she/he/it [-obj?] {V} assist {VS7} perfective ]
6boQ^pu' [ [bare/clipped] {V} assist {VS7} perfective ]
boQ^pu' [ they-them {V} assist {VS7} perfective]
The non-phonetic "feature" symbols in the solution strings are used by various
rules to enforce co-occurrence restrictions and requirements; they can be
ignored by users of the system.
The current system chokes on your heroic bIQong'eghqangchoHmoHlaHpu'neSchugh
because of the way it handles 'egh placement (it is constrained to come after a
transitive verb). There might be other problems, but this one glitch is
sufficient to bring the analysis to a halt. I can fix it pretty easily, but
I'm hoping to get a little more guidance on HOW to fix it. Hence my recent
questions about -moH placement. Your examples are much appreciated.
Another interesting question (with concrete implications for my program) has to
do with noun compounding: What kind of nouns can legally form compounds?
Examples like jolpa' (transport_beam | room) involve just simple noun roots.
The resulting compound can of course take the various nominal suffixes:
[jolpa']mey. But can a noun be suffixed in any way and then continue on to
compound with another noun? We are given ropyaH as "infirmary" (lit. disease
station). Can rop (disease) be augmented before compounding?, as in ?rop'a'yaH
(? critical disease station) or perhaps ?ropHomyaH (?first aid station). I've
allowed this for the time being. Running ropHomyaH I get
rop-Hom|yaH [ disease {NS1} augmentative | duty_station/station ]
rop|Hom|yaH [ disease | bone | duty_station/station ]
In languages that have augmentative and diminutive affixes, some affixed nouns
tend to be become lexicalized: e.g. loD'a' might be translated just as "giant"
rather than as "big man" and loDHom might be translated as "dwarf". They
become almost like simple nouns, and you hardly notice that they are derived.
Even if augmentatives and diminutives cannot compound in general, it would not
be unusual to find the lexicalized ones in compounds.
How about plurals? Can you get something like ?ropmeyyaH (disease plural
station = a diseases station). This is typical of noun compounding in German,
but I can't see any Klingon examples that might suggest it. I've disallowed
such words for now. How about ropqoqyaH ( [apparent disease] station = a
station to treat apparent diseases )?
We are told (p 19) that a compound consists of 2 or 3 nouns in a row. I wonder
if Okrand really intended to set such a limit, but I've made 3 the limit for
now. But I've also put certain apparently lexicalized nouns, like jolpa', in
the lexicon as simple nouns. This allows them to be one of the three elements
of a compound, e.g.
jolpa'SeHlawpaq
jolpa' | SeHlaw | paq [ transport_room | control_panel | book ]
If these words like jolpa' and SeHlaw are not lexicalized, but are just
ordinary compounds placed in the lexicon for convenience, then the example
above is ill-formed (too many compounded nouns).
These are not really points for discussion, unless someone can point me to
examples in the Dictionary that I have missed. Or unless our discussion can
somehow affect the evolution of the language description. With most languages
the thing to do is to march out and corner an informant. I guess Okrand is as
close as we're going to get. Another possibility is to find out what Native
American Languages Okrand worked on as a gradutate student at Berkeley. I
wouldn't be too surprised if he based Klingon on one of them. Then maybe we
could appeal to some real informants.
Best wishes,
Ken Beesley
