LING 567
Lab 4
Chris Curtis and David McHugh (write-up)
04.26.2013

1 & 2.

MATRIX YES-NO QUESTIONS

Matrix clause yes-no questions are indicated in Lakota by the word "he" appearing at the end of the sentence. In formal speech, the word "hwo" can be used in place of "he", but only by male speakers.

Source: b:151
Vetted: s
Judgment: g
Phenomena: q,tam
yaú kta he?
ya-ú ktA he
ú-2SgAgt ktA he
come-2SG.AGT IRR Q
'will you come?'

Source: b:151
Vetted: s
Judgment: g
Phenomena: q,tam,neg
waŋláke šni he?
waŋ-Ø-l-yáŋkA šni he?
waŋyáŋkA-3SgPat-1SgAgt šni he?
see-3SG.PAT-1SG.AGT NEG Q?
'did you not see him?'

Source: author
Vetted: n
Judgment: u
Phenomena: q
yaú he kta?
ya-ú he ktA
ú-2SgAgt he ktA
come-2SG.AGT Q IRR
'will you come?'

We modeled this phenomenon in our grammar by choosing a separate question particle strategy, and setting the position as sentence-final. The form only allows one input, so we chose the more general particle spoken by all speakers: "he". In any case, "hwo" is simply another form of "he" (perhaps at the pragmatic level) and has no syntactic difference.

INFORMATION STRUCTURE

Lakota uses a specific definite article to indicate topic contrast. "k'uŋ" is used in place of the default definite article "kiŋ". These articles appear to have dual roles as both quantifiers and topic markers. Lakota exhibits no overt words or morphemes for focus, however a sentence-initial noun phrase is considered focused. 

Source: b:330
Vetted: s
Judgment: g
Phenomena: info,det
hokšíla kiŋ he hí
hokšíla kiŋ he hí
hokšíla kiŋ he hí
boy DEF.TOP that come_here
'that boy came'

Source: b:330
Vetted: s
Judgment: g
Phenomena: info,det
hokšíla k'uŋ he hí
hokšíla k'uŋ he hí
hokšíla k'uŋ he hí
boy DEF.TOP.CONTR that come_here
'that aforementioned boy came'

To model the topic contrast, we created a new determiner called "def-pst". This determiner's only constraint is that the information-structure meaning is contrastive-topic. Note that we chose this name ("def-pst") to align with how it is most often referenced in gloss--we are not trying to make any claims about tense. The "pst" reference comes from a discourse-level constraint that its complement must have been previously mentioned. As for focus, we attempted to chose clause-initial as the position for focus and included no focus markers. Presumably due to the lack of any marking, this caused our grammar to overgenerate excessively--some sentences generating as many as nine readings.  

COORDINATION

In Lakota, simple coordination ('and') is marked with the lexical item: "na". There is just one coordination strategy, wherein each of the coordinands (excluding the first) is preceded by this "na" item. Some analyses claim that "na" actually follows each of the coordinands (except the last). This analysis is justified by the short pause that's often spoken after "na" when coordinating sentential phrases. However, sources provide no orthographical marking for this pause. Regardless of the analysis, the same strategy is used to coordinate nouns, verbs and sentences alike--however, in the coordination of sentences, there is a variety of vocabulary used to convey such sentiments as "and so" and "and then" for example.
Source: b:374
Vetted: s
Judgment: g
Phenomena: crd
mniȟúha sápiŋ na šá na zí na ská, lená tób úŋpi.
mniȟúha sápA na šá na zí na ská, lená tób úŋ-pi.
mniȟúha sápA na šá na zí na ská, lená tób úŋ-3PlAgt.
cloth black CONJ red CONJ yellow CONJ white these four use-3PL.AGT
'they use cloth of four colors, black, red, yellow and white'

Source: b:374
Vetted: s
Judgment: g
Phenomena: crd
íblotakiŋ na wawáte
í-bl-yotakA na wa-wótA
íyotakA-1SgAgt na wótA-1SgAgt
sit_down-1SG.AGT CONJ eat-1SG.AGT
'I sat down and ate'

Source: author
Vetted: f
Judgment: u
Phenomena: crd
mniȟúha sápiŋ na šá na zí na ská na, lená tób úŋpi.
mniȟúha sápA na šá na zí na ská na, lená tób úŋ-pi.
mniȟúha sápA na šá na zí na ská na, lená tób úŋ-3PlAgt.
cloth black CONJ red CONJ yellow CONJ white CONJ these four use-3PL.AGT
'they use cloth of four colors, black, red, yellow and white'

To model the coordination strategy used in Lakota, we selected one that applies to nouns, NPs, VPs, and sentences. Because all but the first are marked, we chose the markings occur in a polysyndeton fashion by a word spelled "na" that comes before the coordinand. Not only does the questionnaire not provide an option for the alternate analysis wherein "na" follows all but the last coordinand, the reasoning behind this analysis is based on pure prosody.

ATTRIBUTIVE ADJECTIVES

Lakota does not have words that are truly equivalent to English adjectives. Rather, nouns are modified by stative verbs. The construction for adjectives that appear within the noun phrase is simply noun + modifier (stative verb) + determiner. For nouns with an indefinite article, there is an alternate construction that is idiomatically interchangeable with  the standard. This form arranges the words differently and optionally includes the complementizer "čha": noun + indefinite article + modifier (stative verb) + "čha". With certain verbs, "čha" forms a complement clause with a definite article. The "čha" construction's interchangeability with the former provides evidence for the modifying stative verb's being analyzed as a relative clause.

Source: b:626
Vetted: s
Judgment: g
Phenomena: adj
wóyute wašté   uŋyútapi
wóyute wašté   uŋ-yútA-pi
wóyute wašté   yútA-1PlAgt
food   be_good eat-1PL.AGT
'We ate good food.'

Source: author
Vetted: n
Judgment: u
Phenomena: adj
wašté wóyute uŋyútapi
wašté wóyute uŋ-yútA-pi
wašté wóyute yútA-1PlAgt
good food eat-1PL.AGT
'We ate good food.'

Source: author
Vetted: n
Judgment: g
Phenomena: adj
mniȟúha waŋ ská čha opȟéwatȟuŋ
mniȟúha waŋ ská čha opȟé-wa-tȟuŋ
mniȟúha waŋ ská čha opȟétȟuŋ-1PlAgt
cloth be_white IDF C buy-1PL.AGT
'we bought some white cloth'

Source: author
Vetted: n
Judgment: g
Phenomena: adj
mniȟúha waŋ ská opȟéwatȟuŋ
mniȟúha waŋ ská opȟé-wa-tȟuŋ
mniȟúha waŋ ská opȟétȟuŋ-1PlAgt
cloth be_white IDF buy-1PL.AGT
'we bought some white cloth'

This phenomenon has not yet been added to the grammar.

ADVERBS

Adverbs in Lakota can appear either before or after the noun phrase. It is also possible for multiple adverbs to appear in a sentence, and they are free to appear in any sequence. Additionally, adverbs are in some cases used to modify noun phrases. 

Source: b:523
Vetted: s
Judgment: g
Phenomena: adv
huhú kiŋ taŋyéla yasmi
huhú kiŋ taŋyéla yasmi
huhú kiŋ taŋyéla yasmi-3SgAgt
bone DEF well gnaw_clean-3SG.AGT
'he gnawed the bones clean'

Source: b:523
Vetted: s
Judgment: g
Phenomena: adv
taŋyéla huhú kiŋ yasmi
taŋyéla huhú kiŋ yasmi
taŋyéla huhú kiŋ yasmi-3SgAgt
well bone DEF gnaw_clean-3SG.AGT
'he gnawed the bones clean'

Source: b:822
Vetted: s
Judgment: g
Phenomena: adv
wakhéya waŋ skayéla églepi
wakhéya waŋ skayéla égle-pi
wakhéya waŋ skayéla égle-3PlAgt.
tent    IDF whitely put_up_vertical-3PL.AGT
'they put up a white tent'

This phenomenon has not yet been added to the grammar.

EMBEDDED CLAUSES

In Lakota, embedded clauses are marked with a complementizer/relativizer (kiŋ, k'uŋ, or čha). Interrogative clauses use the complementizer "héči", meaning "if" or "whether". The word order between the interrogatives and declaratives in the same. The main verb of the sentence is preceded by the embedded clause.

Source: author, after c:71
Vetted: n
Judgment: g
Phenomena: emb-d
šuŋka kiŋ  aǧúyapi kiŋ yúte kiŋ slolwáye
šuŋka kiŋ  aǧúyapi kiŋ yútA kiŋ slol-wa-y´A
šuŋka kiŋ  aǧúyapi kiŋ yútA kiŋ slolyÁ-1SgAgt
dog DEF bread DEF eat DEF know-1SG.AGT
'I know that the dog ate the bread'

Source: b:814
Vetted: s
Judgment: g
Phenomena: emb-q
héčhamuŋ héči imáyuŋǧe
héčhamuŋ héči i-má-Ø-yuŋǧe
héčhuŋ-3SgPat-1SgAgt héči iyúŋǧA-1SgPat-3SgAgt
do_something-3SG.PAT-1SG.AGT whether ask-1SG.PAT-3SG.AGT
'he asked me if I did it'

Source: author, after b:813
Vetted: n
Judgment: u
Phenomena: emb-d
wičíŋčala kiŋ oȟ'áŋwašte čha slolwáye šni
wičíŋčala kiŋ Ø-oȟ'áŋwašte čha slol-wa-yÁ šni
wičíŋčala kiŋ oȟ'áŋwašte-3SgPat čha slolwÁ-1SgAgt NEG
girl DEF be_kind-3SG.PAT C know-1SG.AGT NEG
'I don't know that the girl is kind'

Source: author
Vetted: u
Judgment: n
Phenomena: emb-d
slolwáye wičíŋčala kiŋ oȟ'áŋwašte čha
slol-wa-yÁ wičíŋčala kiŋ Ø-oȟ'áŋwašte čha
slolwÁ-1SgAgt wičíŋčala kiŋ oȟ'áŋwašte-3SgPat čha
know-1SG.AGT girl DEF be_kind-3SG.PAT C
'I know that the girl is kind'

This phenomenon has not yet been added to the grammar.

NON-VERBAL PREDICATES

Non-verbal predicates do not exist in Lakota. Adjectives, for example, exist only as stative verbs. Nouns can also be used as verbs. The noun for "man," for example, is also used as the stative verb "to be an adult male."

Source: author
Vetted: f
Judgment: g
Phenomena: c
nikáŋ
ni-káŋ
káŋ-2SgPat
be_old-2SG.PAT
'you (sg) are old'

Source author
Vetted: f
Judgment: g
Phenomena: cop
Lakotapi
Lakota-pi
Lakota-3PlPat
be_Lakota-3PL.PAT
'they are Lakota'

This phenomenon does not require any changes to the grammar.

OTHER CHANGES

Other than including the new phenomena and adding vocabulary to the existing verb classes in the lexicon, we made very few modifications to our grammar. We changed from the production version 26 to Sanghoun Song's version 28 to include the information structure features. Another small change we made was the inclusion of the single quote character in the list of punctuation characters to parse, because it is part of the orthography for the newly added word "k'uŋ." Finally we made a tweak to the tdl file manually, to address our auxiliary verb unification issues (described in more detail below).


3. 

Some of the properties not covered by our testsuite were added in previous labs, and still have not been addressed in the grammar. These include the demonstratives and the impersonal verbs (0 valence).

Source: author
Vetted: f
Judgment: g
Phenomena: adj
ká igmú kiŋ wakhúwa
ká igmú kiŋ wa-khúwa
ká igmú kiŋ khuwá-1SgAgt
DEM cat DEF chase-1SG.AGT
'I chase that (yonder) cat'

Source: b:760
Vetted: s
Judgment: g
Phenomena: wo
maǧážu
maǧážu
maǧážu
be_raining
'It is raining'

As described above, several phenomena have been added to the suite, but have not yet been implemented in the grammar. 

The first are the adjectives, which have properties much like stative verbs. More than likely these will be implemented as relative clauses. Our grammar does not support clauses or multiple verbs per sentence (except through coordination or auxiliaries), so these constructions do not parse. Furthermore, the appropriate verb types (nor lexical entries) exist in the lexicon, so the grammar fails to parse on that level as well.

Source: b:626
Vetted: s
Judgment: g
Phenomena: adj
wóyute wašté   uŋyútapi
wóyute wašté   uŋ-yútA-pi
wóyute wašté   yútA-1PlAgt
food   be_good eat-1PL.AGT
'We ate good food.'

Source: author
Vetted: n
Judgment: g
Phenomena: adj
mniȟúha waŋ ská čha opȟéwatȟuŋ
mniȟúha waŋ ská čha opȟé-wa-tȟuŋ
mniȟúha waŋ ská čha opȟétȟuŋ-1PlAgt
cloth be_white IDF C buy-1PL.AGT
'we bought some white cloth'

Examples of adverbs were added to the testsuite, while no changes have yet been made to the grammar for them. Like the adjectives, adverbial lexical types and entries have not been added to the grammar.

Source: b:186
Vetted: s
Judgment: g
Phenomena: adv
ȟ'aŋhíya yápi
ȟ'aŋhíya yÁ-pi
ȟ'aŋhíya yÁ-3PlAgt
slowly  go_there-3PL.AGT
'they went slowly'

Source: b:523
Vetted: s
Judgment: g
Phenomena: adv
huhú kiŋ taŋyéla yasmi
huhú kiŋ taŋyéla yasmi
huhú kiŋ taŋyéla yasmi-3SgAgt
bone DEF well gnaw_clean-3SG.AGT
'he gnawed the bones clean'

Finally, examples of embedded clauses were added to the testsuite. As mentioned above, the grammar does not yet support clauses, so the constructions in the test suite fail to parse. The relativized/complementizers are not all included in the grammar either, as there is overlap in function between them and determiners. As determiners, kiŋ and k'uŋ are in fact in the grammar, but we have not yet implemented their complementizing function.
 
Source: author, after c:71
Vetted: n
Judgment: g
Phenomena: emb-d
šuŋka kiŋ  aǧúyapi kiŋ yúte kiŋ slolwáye
šuŋka kiŋ  aǧúyapi kiŋ yútA kiŋ slol-wa-y´A
šuŋka kiŋ  aǧúyapi kiŋ yútA kiŋ slolyÁ-1SgAgt
dog DEF bread DEF eat DEF know-1SG.AGT
'I know that the dog ate the bread'

Source: b:75
Vetted: s
Judgment: g
Phenomena: emb-d
líla tȟáŋka čha abléze
líla Ø-tȟáŋka čha Ø-ablézA
líla tȟáŋka-3SgPat čha ablézA-3SgAgt
very be_large-3SG.PAT C notice-3SG.AGT
'he noticed that it was very large'

4.

Our grammar yields 72.1% coverage over the test suite, with 3.7% overgeneration, and 40 errors. 10 of the errors are on account of invalid morpheme affix combinations as discussed in the previous lab. The remainder can be attributed to the missing lexicon entries of grammar phenomena not yet implemented. There are 6 well-formed sentences with 0 reading in the grammar--these are also on account of grammar that has not yet been implemented. These 6 are not errors like the rest, as they are composed of just verbs, nouns and determiners. Their constructions according to our grammar so far, however, are invalid.

There are currently two sentences that are reading when they should not. Both of these can be attributed to the auxiliaries sequencing issues. As described below, this is an issue that we improved upon for this lab, but have not completely resolved. It appears as though the AUX verbs are not properly heading their phrases.

We also have one sentence with some ambiguity, as it returns two readings. We think this ambiguity is fine, though. It is a coordination structure between two verb phrases. Since the same strategy applies to sentences, the whole sentence is read as sentence coordination as well. 

Source: b:374
Vetted: s
Judgment: g
Phenomena: crd
íblotakiŋ na wawáte
í-bl-yotakA na wa-wótA
íyotakA-1SgAgt na wótA-1SgAgt
sit_down-1SG.AGT CONJ eat-1SG.AGT
'I sat down and ate'

We interpret this ambiguity as the difference between "I sat down and ate" and "I sat down and I ate." We believe this ambiguity should be expected. 

5.

One of the outstanding issues we had with our grammar was that the customization tool would not generate a grammar that would allow for the chaining of AUX verbs. AUX verb chaining was the way we decided to model the enclitics in our grammar that express aspect, mood, and negation. As it turned out, the customization system was generating AUX verbs inheriting from subj-raise-aux-no-pred, which specifies

  [ ARG-ST < [ ] ,
                   [ LOCAL.CAT.HEAD.AUX - ] > ].

which is what was preventing the chaining. We were, in fact, seeing some limited chaining from our negative AUX, as it was inheriting from subj-raise-aux-with-pred, which does not specify the AUX -. This constraint was not allowing the following sentences to read:

Source: author
Vetted: f
Judgment: u
Phenomena: neg,tam
igmú kiŋ sápa šni kte
igmú kiŋ sápA šni ktA
igmú kiŋ sápA šni ktA
cat DEF be_black NEG IRR
'The cat might/will not be black'

Source: author
Vetted: f
Judgment: u
Phenomena: tam
igmú kiŋ sápa séča kta šni
igmú kiŋ sápA séčA ktA šni
igmú kiŋ sápA séčA ktA šni
cat DEF be_black POSSBL IRR NEG
'The cat probably won't be black'

We adjusted the lakota.tdl file as such:

+subj-raise-aux-no-pred := subj-raise-aux & raise-sem-lex-item. ; &
+;  [ ARG-ST < [ ],
+;             [ LOCAL.CAT.HEAD.AUX - ] > ].

to allow for the desired chaining. Unfortunately, doing so created the repercussion of allowing some enclitic sequencing that should not be allowed. We believe this can be addressed through additional analysis and refactoring of the FORM features on those auxiliary verbs.
