Re: Commentary: From Tokana to Denden (LONG)

This message doesn't seem to have gone through, so I'm sending it
again.  Apologies if you get two copies!  -- Matt.


>I wrote:
>
>>> I must not have provided nearly enough info, because you seem to have
>>> parsed these sentences incorrectly.  The correct structure is:
>
>And Boudewijn Rempt replied:
>
>>Well, what made me go wrong was the bit in the notes:
>>
>> So "itan maha olat" in the above poem can only mean "that which
>> one hears". It can't mean "that which hears things".
>>
>>So I took that for the relative clause, and tried to connect the
>>dative 'suhoi' and 'sihe' and so on to the verb.
>
>To which I reply:
>
>Clumsy me.  I should have spelled out what I meant in more detail.
>Perhaps if I'd phrased it this way:  <quote> "Itan maha olat..." can
>only mean "That which is heard...".  It can't mean "That which
>hears...".  <end quote>
>
>Then again, maybe that isn't all that much clearer (see below).
>
>I wrote:
>
>>>         [ itan maha olat suhoi ]  uthma tsampatin
>>>         [ that which (one) hears in-the-wind ]  gives health
>>>
>>>         [ itan maha olat sihe ]  uthma pamihati
>>>         [ that which (one) hears in-the-stream ]  gives nourishment
>>>
>>> So SUHOI and SIHE are part of the relative clause.
>>> The ending -AT on OLAT indicates that this verb is part of an embedded
>>> clause (here, a relative clause headed by MAHA "what").  The ending -A on
>>> UTHMA, by contrast, indicates that this verb is in the main clause.
>
>And Boudewijn replied:
>
>>Well, I'd really like to see a thorough syntactic analysis of this
>>constructions - it's still not entirely clear to me.
>
>OK, you asked for it!  Let's begin with the basics:
>
>(1)  Noun phrases in Tokana are inflected for case.  The absolutive and
>ergative
>(or nominative) cases are unmarked, while the remaining cases (dative,
>instrumental, ablative, and comitative) are marked by suffixes on the noun.
>
>(2)  Definite noun phrases are preceded by a determiner which 'agrees' in case
>with the noun (like in German).  When the noun phrase in question is an
>argument
>of the verb and appears right-adjacent to the verb, then the determiner
>attaches
>to the verb as a suffix (cf. "-n" and "-na" in (3) below).
>
>(3)  In main clauses, unmarked order is SVO or OVS:
>
>       Na          moiha   kahta-n         mikal
>       the.ERG   girl      hit-the.ABS   boy
>       "The girl hits the boy"
>
>       Ne           mikal   kahta-na       moiha
>       the.ABS   boy      hit-the.ERG   girl
>       "The girl hits the boy"
>
>The OVS (or 'inverse') sentence is often given a passive-type reading:
>"The boy
>is hit by the girl".
>
>(4)  Transitive sentences like the ones in (3) contain an ergative subject and
>an absolutive object.  Either of these arguments may be deleted if its
>referent
>is unknown or irrelevant.  When the absolutive object is deleted, a special
>'antipassive' prefix "u-" must be added to the verb.  The word order is SV:
>
>       Na           moiha   u-kahta
>       the.ERG   girl     ANTIPASS-hit
>       "The girl hits [things]"
>       "The girl is a hitter"
>
>The ergative subject, on the other hand, can be deleted freely, without the
>need for a special marker on the verb.  When the subject is absent, the word
>order is OV:
>
>       Ne          mikal   kahta
>       the.ABS   boy    hit
>       "The boy is hit"
>       "[Someone] hits the boy"
>
>(5)  In embedded clauses, the word order is typically VSO (VOS is also
>possible, but less frequent).  In order to indicate that the clause is
>embedded,
>one of a special set of 'dependent order' suffixes is added to the verb.
>This
>suffix appears after the verb stem and before the suffixed determiner (if
>any):
>
>       ...   kahta-n-na           moiha   ne          mikal
>       ...   hit-DEP-the.ERG   girl      the.ABS   boy
>       ...   "that the girl hits the boy"
>
>For example:
>
>       Ikime      iona     kahta-n-na           moiha   ne         mikal
>       we.DAT   know   hit-DEP-the.ERG   girl     the.ABS   boy
>       "We know that the girl hits the boy"
>
>(6)  Just as in main clauses, embedded clause arguments may be freely
>deleted if their referents are unknown or irrelevant.  As in main clauses,
>deletion of the absolutive object is accompanied by the prefixing of "u-"
>to the verb, while deletion of the ergative subject requires no verb
>marking.  The word order is VS or VO:
>
>       Ikime       iona    u-kahta-n-na                         moiha
>       we.DAT   know   ANTIPASS-hit-DEP-the.ERG   girl
>       "We know that the girl hits [things]"
>
>       Ikime      iona     kahta-n-ne           mikal
>       we.DAT   know   hit-DEP-the.ABS   boy
>       "We know that [someone] hits the boy"
>       "We know that the boy is hit"
>
>(7)  Relative clauses are formed using dependent suffixes on the verb
>as well.  In relative clauses, one of the arguments (viz. the one
>corresponding to the head of the relative clause) is obligatorily absent.
>Relative clauses always follow the noun they modify, just like in English:
>
>       moiha   [ kahta-n-ne          mikal ]
>       girl         hit-DEP-the.ABS   boy
>       "a girl who hits the boy"
>
>Here, the relative clause "kahtanne mikal" modifies "moiha".  Notice
>that the argument inside the relative clause is marked with absolutive
>case.  Thus, we know that it is being interpreted as a direct object,
>which means that the missing argument must be the subject.  This is
>how we know that the girl is the one who is doing the hitting.
>
>       mikal   [ kahta-n-na          moiha ]
>       boy        hit-DEP-the.ERG   girl
>       "a boy whom the girl hits"
>       "a boy who is hit by the girl"
>
>Here, the relative clause "kahtanna moiha" modifies "mikal".  Notice
>that the argument inside the relative clause is marked with ergative
>case, and is therefore the subject of the verb "hit".  This is how we
>know that the boy is the one being hit.
>
>(8)  In relative clauses, just as in other kinds of embedded clauses,
>it is possible to delete an argument whose referent is unknown or
>irrelevant.  As before, a deleted absolutive is marked by the prefix
>"u-", while a deleted ergative is unmarked:
>
>       moiha   [ u-kahta-t ]
>       girl         ANTIPASS-hit-DEP
>       "a girl who hits [things]"
>       "a girl who is a hitter"
>
>(Notice that we have "-t" instead of "-n" as the dependent suffix in
>this case.  "-t" and "-n" are just phonological variants of the same
>suffix.)
>
>       mikal   [ kahta-t ]
>       boy        hit-DEP
>       "a boy who is hit"
>
>Notice that in both expressions, "moiha ukahtat" and "mikal kahtat",
>there is no subject or object after the verb.  In other words, these
>relative clauses contain 2 gaps.  Because the verb is prefixed with "u-"
>in the first case, we know that the verb denotes an action directed at
>an unknown/irrelevant object; therefore the noun being modified by
>the relative clause ("girl") must correspond to the subject of the verb.
>
>In the second case ("mikal kahtat") there is no "u-" prefix on the verb.
>Therefore we know that the direct object is NOT unknown/irrelevant.
>Therefore, it must be the SUBJECT which is unknown/irrelevant, while
>the missing object corresponds to the head of the relative clause ("boy").
>
>(9)  It is also possible to form what are called 'free relative' construc-
>tions, equivalent to the following:
>
>       the one who hits the boy
>       s/he who hits the boy
>
>       the thing which the girl hits
>       that which is hit by the girl
>       what the girl hits
>
>These are formed using a series of interrogative/indefinite
>quantifiers roughly equivalent to "wh-" words in English:
>
>       mioh    "who, someone"
>       mah     "what, something"
>       melh    "where, somewhere"
>       ima     "when, sometime"   etc.
>
>If a determiner is added to these, they become definite expressions:
>
>       ne mioh "the one who...", "s/he who..."
>       te mah  "the thing which...", "that which..."
>       te melh "the place where..."
>       te ima  "the time when..."
>
>These are literally "the who", "the what", "the where", and "the
>when".  (Note that "ne" is the unmarked determiner used with
>animates, while "te" is the unmarked determiner for inanimates.)
>
>Putting the determiner + interrogative together with an embedded
>relative clause gets us the following:
>
>       ne mioh   [ kahta-n-ne           mikal ]
>       the who     hit-DEP-the.ABS   boy
>       "the one who hits the boy"
>       "s/he who hits the boy"
>
>       ne mioh   [ u-kahta-t  ]
>       the who     ANTIPASS-hit-DEP
>       "the one who hits [things]"
>
>       ne mioh   [ kahta-n-na         moiha ]
>       the who     hit-DEP-the.ERG  girl
>       "the one who is hit by the girl"
>
>       ne mioh   [ kahta-t   ]
>       the who     hit-DEP
>       "the one who is hit [by someone]"
>
>Whew!!!!
>
>So, the subject noun phrase in the Tokana poem has the same form as
>this last example:
>
>       te mah      [ ola-t         suhoi        ]
>       the what     hear-DEP   wind.DAT
>       "the thing which is heard [by someone] in the wind"
>       "that which is heard in the wind"
>
>Now, to put this phrase inside the larger sentence:
>
>The verb phrase is:
>
>       uthma   tsampatin
>       give      body-balance
>       "gives balance to the body"
>
>Since the word order is SVO, the noun phrase "te mah olat suhoi" will
>precede the verb.  One final wrinkle:  Since this expression is the
>subject of the sentence, it must appear in the appropriate case.  When
>a subject is animate, the ergative case is normally used (as in
>the sentences about the girl hitting the boy).  However, "te mah olat
>suhoi" is inanimate, and inanimate subjects must appear in the
>instrumental case.  Thus, the proper form will be "itan maha olat
>suhoi" (where "itan" is the instrumental form of the inanimate
>determiner "te", and "-a" is the instrumental noun suffix):
>
>[ Itan         maha          [ ola-t         suhoi        ] ]   uthma
>tsampatin
>   the.INST what.INST     hear-DEP  wind.DAT        give      body-balance
>"That which is heard in the wind gives balance to the body"
>
>Is that clear at last?  I hope so...
>
>Matt.
>
>