Minhast/Noun Incorporation: Difference between revisions
(35 intermediate revisions by the same user not shown) | |||
Line 1: | Line 1: | ||
{{Construction}} | |||
= Introduction = | = Introduction = | ||
= Types of Noun Incorporation = | = Types of Noun Incorporation = | ||
== Mithun's Framework == | == Mithun's Framework == | ||
Languages which employ NI do have alternative, analytic structures containing the same semantic information as an NI structure does. However, NI provides a way to manipulate discourse, reduce the salience of an entity in order that other entities can take precedence in extended speech, provide stylistic and rhetorical alternatives to their corresponding analytic expressions, and even derive new lexical items. Mithun (1984) has identified four categories of NI that occur cross-linguistically. | Languages which employ NI do have alternative, analytic structures containing the same semantic information as an NI structure does. However, NI provides a way to manipulate discourse, reduce the salience of an entity in order that other entities can take precedence in extended speech, provide stylistic and rhetorical alternatives to their corresponding analytic expressions, and even derive new lexical items. Mithun (1984) has identified four categories of NI that occur cross-linguistically. The four categories Mithun has identified are: | ||
# Type I - Lexical Compounding: the creation of new lexical items by compounding a noun root and verb root; | |||
# Type II - Case Manipulation: a noun (usually a Patient, although Instrumental and Locative nouns may be involved) is incorporated into the verb complex. This is a valence operation: if the incorporated noun was originally a core argument, another argument can occupy the position vacated by the IN and assume core status. Alternatively, depending on the semantic nature of the verb, Oblique8 nouns that are Instruments, Locatives, or Goals may also be incorporated;9 | |||
# Type III - Manipulation of Discourse: NI is used to background information in sections of the discourse so that other arguments are brought to the foreground. It allows speech participants to focus on the important entities within a particular passage of the discourse; | |||
# Type IV - Classificatory NI: Mithun describes this type of NI wherein a “...relatively general N(oun) stem is incorporated to narrow the scope of the V(erb)...but the compound noun stem can be accompanied by a more specific external NP which identifies the argument implied by the IN.” | |||
Not all languages that employ NI exhibit all four types. For example, in the indigenous Brazilian language Karajá, NI is mostly restricted inalienably possessed nouns, particularly body parts, and does not alter valence, which by nature involves Type II case manipulation (Ribeiro 2001). Chukchi freely employs Type II NI to alter argument structure, Type III is attested per Mithun, but Type IV NI is absent (Modena & Muro). Nevertheless, languages that employ all four types of NI are found in disparate languages, including Mohawk, Caddo, and Gunwinggu. Minhast also falls within this group. | |||
=== Type I Noun Incorporation - Lexical Compound Formation === | |||
Minhast utilizes Type I NI to create lexical compounds but only if a particular activity, state, or event occurs frequently to warrant institutionalizing into the lexicon. Usually, one or both elements of the compound are shortened, as in the following examples: | |||
5a) ''aydann''- “To store water in a cistern, reservoir, or catch-basin” (derived from ''ayaya'' - “put something into a container” + ''dannua'' “water”)<br/> | |||
5b) ''uzdann- → uzzat-dannua'' “To draw water from a well” (derived from ''uzzat''- “to pull | |||
something out of a container or other enclosing object, e.g. an envelope” + dannua “water” | |||
Compounding NI is a derivational process. If the compound yields a new verb, it has the fullstatus and capabilities of a verb not derived from compounding, including NI:<br/> | |||
5c) ''Aydantayattaran → aydann-tayatta-ar-an'' “He poisoned the well” (lit. “He stored the water with poison). | |||
=== Type II Noun Incorporation - Case Manipulation === | |||
As Mithun identified in Case Manipulation NI, an important function of IN Minhast is to alter the argument structure of a clause. The prototypical function of NI is to decrease the valency of a verb; the transitivity of a clause is decreased by removing one of the core arguments, namely the PT, and absorbing it into the verb. This opens up the Absolutive position to be occupied by another argument, either an oblique argument, or the Ergative argument. | |||
Returning back to Sentence 1a and 1b, the argument structure has been altered from a transitive clause in Sentence 1a to an intransitive one via the application of NI previously observed in Sentence 1b. Both sentences have been restated here as Sentences 5a and 5b: | |||
5a) Yadukte kaslub ayupparu → yaduk=de kaslub ayupp-ar-u (boy=ERG dog.ABS point.at-PST TRANS) | |||
“The boy pointed at the dog.” | |||
5b) Yaduk ayukkaslubaran → yaduk ayup-kaslub-ar-an (boy.ABS point.at-dog-PST- INTRANS) | |||
“The boy pointed at a/the dog” (lit: “The boy dog-pointed”). | |||
The alteration of the argument structure from a transitive sentence to an intransitive one is pragmatically motivated and changes the nature of the discourse. The incorporation of the PT kaslub, both opens up the Absolutive position for occupation by another argument, in this case, the Agent yaduk. The incorporation of the Patient kaslub also backgrounds it, reducing its salience in the discourse. The Agent yaduk thus becomes more salient, as it has now become the sole core argument of the sentence. The result alters discourse by presenting the Agent as the most important element of the discourse, while that of the Patient has been reduced to a peripheral role. | |||
Patients are not the only arguments that can be subjected to NI. An interesting feature of NI in Minhast is that the semantic nature of a verb may allow certain non-PTs oblique arguments, namely Instrumentals and Locatives, to be incorporated. This is similar to Ainu, another polysynthetic language of Northeast Asia unrelated to Minhast . Like Minhast, Ainu can target Instrumental arguments for NI, like sapa (“head”) underlined below in Sentence 5: | |||
13 This is an instance of Case Manipulation NI, i.e. Mithun's Type II classification. | |||
14 There still remain some die-hard advocates who group the two languages with Chutchki and Nivkh under a single | |||
family called Amuran, a reference to the Amur Riverine System. Surface similarities have been ascribed to areal features. | |||
6) Ratki apa a-sapa-e-puni → hung door 1s.A-head-APPL-lift | |||
“I lifted the suspended door with my head.” (J. Runner & Raul Aranovich 2003). | |||
NI of Oblique arguments in transitive clauses does not affect valency, since the Absolutive position remains occupied by the PT. However, the incorporation of Instrumental and Locative arguments modify the meaning of the verb. Speakers are aware that in utilizing this form of NI, they are essentially creating new vocabulary on-the-fly. These new words may be created as one-time entities for the current speech event, or they may be institutionalized and become permanent vocabulary in the lexicon. Examples of Instrumental and Oblique NI are demonstrated in Sentences 6a – 6e: | |||
7a) Yakte dūy kallutaššiakaru → yak=de dūy kallut-haššia-ak-ar-u (1S=ERG salmon.ABS eat- with.chopsticks-3P.INANIM.ABS+1S.ERG-PST-TRANS) – lit: “I eat.with.chopsticks the salmon.” | |||
7b) Yak (dūyaran) kallutaššiekarampi → yak (dūy=aran) kallut-haššia-ek-ar-an-pi (1S.ABS (salmon=DAT) eat.with.chopsticks-1S.ABS-PST-INTRANS-ANTI) | |||
“I eat (some salmon) with chopsticks.” | |||
7c) Yak asunkallutaššiekaran → yak kallut-haššia-ek-ar-an (1S.ABS HAB-eat.with.chopsticks- 1S.ABS-PST-INTRANS) | |||
“I eat/am eating with chopsticks.” | |||
7d) Yak iknatūmanekaran → yak ikna-tūman-ek-ar-an (1S.ABS go-house-1S.ABS-PST- INTRANS) | |||
“I went home.” | |||
7e) Yak ikassuhūrekaran → yak ikassu-hūr-ek-ar-an (1S.ABS rest-mountain-1S.ABS-PST- INTRANS) | |||
“I rested on the mountain.” | |||
The surfacing of the Antipassive in Sentence 7b illustrates that NI of Instrumentals, in this case haššia, does not and cannot decrease valency. Instead, Antipassivation is the only valency-decreasing operation available to reduce the valency of a verb whenever an Instrumental oblique argument is incorporated. | |||
Similarly, the incorporation of a Locative or Goal noun into a positional or locomotive verb does not affect valency, as in Sentence 7d and 7e. | |||
Sentence 7b illustrates that no discussion of Noun Incorporation in Minhast can be made without explaining the interrelation between NI and Antipassivation. Some polysynthetic languages use both NI and Antipassivation to manipulate argument structures. Minhast is among this group of languages. An observation noted among these languages with both NI and Antipassivation is that NI automatically triggers the Antipassive to surface, as demonstrated in Sentences 8a and 8b, taken from Nishga, a North American indigenous tribe of the Pacific Northwest (Mithun, 1984); and Sentences 8c and 8d, taken from Yucatec Mayan (Bricker, 1978): <br/><br/> | |||
8a) səməyé:n-sk -m-hó:n → simijeehisgumhoon to.smoke-ANTIPASS-ADJ-fish 'To smoke fish'<br/> | |||
8b) íc'l-'sk -m-tá:la → lits'ilsgumdaala to. count. up-ANTIPASS-ADJ-money 'to keep track of | |||
money donated at a potlatch'<br/> | |||
8c) t-in-č'ak-∅-ah če'. COMP-I-chop-it-PERF tree “I chopped a tree.”<br/> | |||
8d) č'ak-če'-n-ah-en. Chop-tree-ANTIPASS-PERF-I(ABS) “I wood-chopped” = “I chopped wood.”<br/> | |||
There are other languages that utilize both NI and Antipassivation where NI does not trigger the Antipassive to surface. The Northeast Asian language Chukchi, a branch of the Chukotko-Kamchatkan is such an example (Kozinsky 1976). Sentence 9a shows NI of an Instrumental noun without Antipassivation. However, Sentence 9b the Antipassive affix appears, but NI is absent because the Antipassive has demoted the PT qora to an oblique Instrumental argument. | |||
9a) Kopalhin na-aldt-koqenat. → walrus.blubber they-knife-mincing “They are mincing walrus blubber with a knife. | |||
9b) Ine-lqerir-ə-rkən (qora-ta). → APASS-seek-PRES-/3sS (deer-INSTR) “He is seeking (for deer).” | |||
These two examples demonstrate that NI and Antipassivation are two distinct processes in Chukchi, whereas in Yucatec Mayan they appear to be a unitary process, as they appear simultaneously. | |||
Like Chukchi, Minhast treats NI and Antipassivation as two separate, distinct processes. In Minhast, NI and Antipassivation may complement each other, and other times they may be mutually incompatible. Sentences 10a and 10b are presented below. Sentence 10a shows a situation where both NI of an Instrument and the surfacing of the Antipassive affix -pi- occur in the same sentence: | |||
10a) Yak dūyaran kallutaššiekaranampamā, [PRO] yusnakekaran, wassetta. | |||
→ yak dūy=aran kallut-haššia-ek-ar-an-pi=mā, [PRO] yusnak-ek-ar-an, wa=setta (1S.ABS | |||
salmon=DAT eat.with.chopsticks-1S.ABS-PST-INTRANS-ANTI=SUBORD, [PRO] be.late- 1S.ABS-PST-INTRANS, CONN=even.though) | |||
“I ate (some) salmon with chopsticks, even though I was late.” | |||
As mentioned earlier, incorporation of an Instrumental noun does not decrease valence, so Antipassivation was required to decrease the valency of Sentence 10a. Here, an Instrumental noun, haššia, has already been incorporated into the verb kallut (to eat). Since only one noun can be incorporated into a verb complex at any one time, the only way to reduce the valency of the clause is to apply Antipassivation to demote the semantic Patient dūy (salmon) from core status. The application of | |||
15 The situation with Yucatec Mayan is somewhat more complicated. The language displays split ergativity conditioned by tense: Erg-Abs alignment is observed in the perfect aspect (glossed by Bricker as COMP “completive”), while Nom- Acc alignment is found in the imperfect aspect. This is why the Antipassive does not surface when NI occurs in the imperfect aspect, because Antipassives occur only in the Erg-Abs component of split ergative systems. | |||
the Antipassive marker -pi has demoted dūy from core Absolutive status to a Dative oblique argument marked with the clitic =aran, the typical oblique marker used to mark demoted semantic Patients. Such an operation is required if the speaker wishes to convert yak to Absolutive status for syntactic reasons (e.g. to become the S/O pivot16 to feed into subordinate clauses, as in this case). | |||
In contrast, the situation in Sentence 10b shows NI of a Patient. Here, the NI of the Patient blocks Antipassivation: | |||
10b) Yaduk ayukkaslubaran → yaduk ayup-kaslub-ar-an (boy.ABS point.at-dog-PST- INTRANS) – “The boy pointed at a/the dog” (lit: “The boy dog-pointed”). | |||
Here, the semantic Patient kaslub (dog) has been incorporated into the verb ayup (to point), but here the Antipassive does not surface. Nor can it without rendering the sentence ungrammatical, because Antipassivation reconfigures a sentence's argument structure by demoting or removing Patients. When kaslub was incorporated, there was no Patient available for the Antipassive to target because the Patient had already been removed from core status by NI. | |||
To illustrate the reason why Antipassivation was blocked by NI of a Patient, Sentence 10c presents a theoretical situation resulting from the application of Antipassivation after a Patient has been incorporated: | |||
10c) **Yadukaran [missing ABS] ayukkaslubarampi → yaduk=aran [missing ABS] ayup- kaslub-ar-an-pi (boy.ABS point.at-dog-PST-INTRANS-ANTI) | |||
Sentence 10c is ungrammatical because NI has already removed the PT (kaslub), resulting in a | |||
monovalent clause with only one core argument, the Absolutive (yaduk). Antipassivation would | |||
Sap min Aynuw-ast min gubbāturrād sattabe- | |||
demote the Absolutive yaduk to a Dative peripheral argument, yadukaran, leaving no other argument to 17 | |||
fill in the missing Absolutive position. Minhast forbids zero-valent sentences , so the simultaneous application of NI and Antipassivation renders Sentence 10c infelicitous. | |||
Thus, both Minhast and Chukchi treat NI and Antipassivation as two separate, distinct operations. But whereas the Chukchi samples showed that the Antipassive does not surface when even an Instrumental noun is incorporated, Minhast allows Antipassivation of Instruments if valency reduction to demote a Patient is required, allowable provided the IN is an Instrumental argument. | |||
16 A pivot is a syntactic construct in multi-clause sentences wherein the syntactic role of a core argument that has been omitted by PRO-drop, is recovered from its role in a higher level clause. There are two types of Pivots identified by Robert Dixon in Ergativity (1994): S/A Pivots and S/O Pivots. S/A Pivots, associated with Nominative-Accusative language, assign the Nominative argument as the argument (Pivot) that is coreferrent with the null argument of successive clauses. Syntactically Ergative languages, in contrast, assign the Absolutive as the Pivot of successive clauses. | |||
17 There is one exception to this rule: Minhast has a closed set of Interrogative verbs, some of which take zero core arguments. | |||
-ar-an- | |||
=naft yak=de nusill-ek-u, ...[PRO | |||
= gubbāturrād] nurr-wastane-0-ek-nes-u-d | |||
“Sap min Aynuw-ast min gubbāturrād | |||
sattabe-redad-ar-an=naft yak=de nusill-ek-u” | |||
The use of Case Manipulation accounts for the majority of NI in Minhast. This is not surprising, since Minhast, as a syntactically ergative language, utilizes various grammatical devices to maintain and manipulate the S/O pivot to cross-reference the Absolutive argument across clauses. | |||
=== Type III Noun Incorporation - Discourse Manipulation === | |||
Another important difference in Minhast between Antipassivation and Noun Incorporation is found in discourse manipulation in extended speech or narratives. Antipassivation is often used to remove an element from discourse entirely, marking the demoted NP as truly incidental and ultimately unimportant to the narrative. On the other hand, Minhast uses NI to retain the demoted argument in the | |||
18 | |||
discourse, albeit backgrounded . Backgrounding is a useful device if the salience of the demoted | |||
argument must be decreased, but is still too important to eliminate from the discourse. Therefore, an IN may be retained throughout the length of the narrative. In doing so, the speaker can maintain the context in which the core arguments operate, a useful mnemonic and rhetorical tool to keep the listener | |||
19 | |||
on track with the discourse . Additionally, the IN may be promoted back to a core argument later on | |||
without creating an abrupt discontinuity within the narrative, which would occur if an argument that had been supposedly marked incidental in the discourse by Antipassivation were to be reintroduced into the argument. Indeed, it is the case that speakers often expect a noun backgrounded by NI in extended discourse to later resume core status at a later point. The following sample passage comes from a recording of a Salmon Speaker conversation, and illustrates the backgrounding function of NI: | |||
11) Ayakna ubarrāsekabampamā, dūyaran ehahekampi, kašmakki ehaheku, arruššan sakwaštan, yakmaran saxtihayyuštinesaš, saxtimurrešnesašnaft irraktuyeku, wahēk intušindūyekan, ayayadūyekan, indiknatūmanekammā, hūraran indiħyekan, reyriaktipār sipsambuyekammā, iyuššitaran sarekampamā, intadutyusapakmahummā, dumyaharan sarekampi, yapkekan, uzzattuyekan, bastettuyekanimmā, isangarekan. Wahēk, uzzatekunimmā, inditayyisakšareku, alluakan. | |||
Gloss: Ayakna ubarrās-ek-ab-an-pi=mā dūy=aran ehah-ek-an-pi, kašmak=ki ehah-ek-u, arrusi-š-an sakwat-š-an, yakma=ran saxt-hayyut-š-nes-an=š, saxt-murre-š-nes-an=š=naft irrak-dūy-ek-u, wahēk nta-ušn-dūy-ek-an, ayaya-dūy-ek-an, nd-ikna-tūman-ek-an=mā, hūr=aran nd-hy-ek-an reyria-k-de=pār sipsab-dūy-ek-an=mā, iyuššit=aran sar-ek-an-pi, nta- dut-yusap-akmah-u=mā, dumyaħ=aran sar-ek-an-pi, yapk-ek-an, uzzat-dūy-ek-an, bastet-dūy- ek-an-nimmā, isangar-ek-an. Wahēk, uzzat-ek-u-nimmā, n-tayy-sakšar-ek-u, alluak-an. | |||
early.morning to.fish-1S.ABS-IMPF-INTRANS-ANTI=SUBORD salmon=DAT capture- 1S.ABS-INTRANS-ANTI, kašmak.river=LOC capture-1S.ERG-TRANS, be.big-ANIM- INTRANS be.fat-ANIM-INTRANS, salmon.noodle.soup=DAT INCHO-dinner-ANIM.ABS- IMMED.FUT-INTRANS=IRREAL INCHO-be.delicious-ANIM-IMMED.FUT- INTRANS=IRREAL=NMLZ know.general.truth-salmon-3S.COMM.ABS + 1S.ERG-TRANS, wahēk INTENS-hit-salmon-1S.ABS.INTRANS, INCEP-go-home-1S.ABS- INTRANS=SUBORD mountain=DAT INCEP-be.high-1S.ABS-INTRANS shoulder-1S.ERG- ERG=INSTR carry-salmon-1S.ABS-INTRANS=SUBORD, storm.clouds= DAT see-1S.ABS- INTRANS-ANTI INTENS-DAT.APPL-rain-1S.ABS+3P.ERG-TRANS=SUBORD, cave=DAT see-1S.ABS-INTRANS-ANTI, take.shelter.from.bad.weather-1S.ABS-INTRANS, pull.out.of.container-salmon-1S.ABS-INTRANS, place.down.carefully-salmon-1S.ABS- INTRANS=PURPOS.SUBORD light.up.a.fire-1S.ABS-INTRANS . wahēk, | |||
18 Discourse Manipulation falls under Mithun's Type III classification. But the discourse manipulation function of backgrounding information in Minhast is very different from other languages with NI, especially the oft-cited works on Mohawk NI processes. These differences are described in greater detail later in the text. | |||
19 The use of NI to background previously mentioned items falls under Mithun's Type III Noun Incorporation category. | |||
pull.out.of.container-3S.COMMON.ABS+1S.ERG-TRANS-PURPOS.SUBORD INCEP- prepare.food-REFL.BEN-3S.ABS+1S.ERG-TRANS, be.pregnant.with.eggs-INTRANS. | |||
“Early in the morning, I was fishing and caught a salmon in the Kashmak River. It was a big, fat one, and I knew it would make good yakma for dinner. So I killed it and put it in my basket, and I began my journey home, heading up the mountain carrying the salmon over my shoulder, but then I saw storm-clouds and soon the rains began to pelt me. I found a small cave and took shelter in it, and then I took out the salmon and lay it aside so I could make a fire. I took the salmon out and began to clean it, it was a female and filled with roe.” | |||
The passage introduces the salmon in the beginning: “...I was fishing and caught a salmon...” Here, the salmon, while semantically Patient, is introduced in the narrative as a Dative argument, and becomes the topic of interest in the next few clauses, where it assumes core status as the Absolutive argument. Casting the salmon as an Absolutive argument also ensures that the sentence fulfills the requirements for its Absolutive argument to become the S/O pivot of the next clause, which is a stative clause. Minhast lacks a separate word class for adjectives, hence stative verbs are used instead. Statives target only Absolutives, never Ergatives or Obliques. After the salmon is killed and placed in the basket, the wahēk particle20 allows the narrator to resume his status as the Absolutive argument so that he can serve as the S/O pivot for the next series of clauses up to the point in the narrative where he makes a fire. Throughout this part of the narrative, the salmon, when mentioned, has been backgrounded by NI so that the narrator can retain his status as the S/O pivot. The second wahēk of the narrative in the next clause allows the narrator to change the S/O pivot again, at which point he reassigns the salmon to Absolutive status in order for it to become the S/O pivot to feed the final three clauses of the narrative. | |||
The same native speaker who provided this narrative was presented an alternative text, whose English translation is exactly the same, but instead of NI to demote the salmon and move it into the background, Antipassivation was used instead to demote the salmon to a Dative oblique argument: | |||
12) Ayakna ubarrāsekabampamā dūyaran ehahekampi, kašmakki ehaheku, arruššan sakwaštan, yakmaran saxtihayyuštinesaš saxtimurrešnesašnaft irakeku, wahēk dūyaran intušnekampi, dūyaran ayayakampi, indiknatūmanekammā hūraran indiħyekan reyriaktipār dūyaran sipsabekampamā, iyuššitaran sarekampamā intadutyusapakmahummā, dumyaharan sarekampi, yapkekan, dūyaran uzzatekampi, dūyaran bastetekampinimmā isangarekan. Uzzatekunimmā inditayyisakšareku, alluakan. | |||
Gloss: Ayakna ubarrās-ek-ab-an-pi=mā dūy=aran ehah-ek-an-pi, kašmak=ki ehah-ek-u, arrusi-š-an sakwat-š-an, yakma=aran saxt-hayyu-š-nes-an=š saxt-murre-š-nes-an=š=naft irak- ek-u, wahēk dūy=aran nt=ušn=ek=an-pi, dūy=aran ayaya-ek-an-pi, nd-ikna-tūman-ek-an=mā hūr=aran nd-ħy-ek-an reyria-ek=de=pār dūy=aran sipsab-ek-an-pi=mā, iyuššit=aran sar-ek- an-pi=mā nta-dut-yusap-s-ek-u=mā, dumyah=aran sar-ek-an-pi, yapk-ek-an, dūy=aran uzzat- ek-an-pi, dūy=aran bastet-ek-an-pi=nimmā isangar-ek-an. Uzzat-ek-u=nimmā nd-tayy-sakšar- ek-u, alluak-an. | |||
early.morning to.fish-1S.ABS-IMPF-INTRANS-ANTI=SUBORD salmon=DAT capture- 1S.ABS-INTRANS-ANTI, kašmak.river=LOC capture-1S.ERG-TRANS, be.big-ANIM- | |||
20 The wahēk particle performs multiple functions. It is a type of subordinative sequential particle roughly translated as “and then”. It differs from the -mā suffix in that it simultaneously allows the Absolutive and Ergative arguments to switch roles. The switch in roles between the core arguments coincides with a change in topic, so wahēk serves as a topic shifter as well. | |||
INTRANS be.fat-ANIM-INTRANS, salmon.noodle.soup=DAT INCHO-dinner-ANIM.ABS- IMMED.FUT-INTRANS=IRREAL INCHO-be.delicious-ANIM-IMMED.FUT- INTRANS=IRREAL=NMLZ salmon=DAT know.general.truth-3S.COMM.ABS + 1S.ERG- TRANS, wahēk salmon=DAT INTENS-hit-1S.ABS.INTRANS-ANTI, INCEP-go-home- 1S.ABS-INTRANS=SUBORD mountain=DAT INCEP-be.high-1S.ABS-INTRANS shoulder- 1S.ERG-ERG=INSTR salmon=DAT carry-1S.ABS-INTRANS-ANTI=SUBORD, storm.clouds= DAT see-1S.ABS-INTRANS-ANTI INTENS-DAT.APPL-rain- 1S.ABS+3P.ERG-TRANS=SUBORD, cave=DAT see-1S.ABS-INTRANS-ANTI, take.shelter.from.bad.weather-1S.ABS-INTRANS, salmon=DAT pull.out.of.container-1S.ABS- INTRANS-ANTI, salmon=DAT place.down.carefully-1S.ABS-INTRANS- ANTI=PURPOS.SUBORD light.up.a.fire-1S.ABS-INTRANS. wahēk, pull.out.of.container- 3S.COMMON.ABS+1S.ERG-TRANS-PURPOS.SUBORD INCEP-prepare.food-REFL.BEN- 3S.ABS+1S.ERG-TRANS, be.pregnant.with.eggs-INTRANS. | |||
When asked about the alternate narrative, the consultant said that while it was grammatically well- formed, he had not expected to hear it mentioned multiple times throughout the narrative. He remarked that the salmon had been demoted by the first occurrence of the Antipassive, and asked why the Antipassive was used if the salmon was important enough to remain a part of the narrative until it would later resume its core status. | |||
In this respect, NI in Minhast differs in its discourse function from that of Mohawk. An IN in extended Mohawk discourse, as backgrounded information, is truly incidental information, meant only to provide contextual information in the body of the discourse, as illustrated in this passage (Mithun:1984): | |||
13) ó:nv yeyóhe n-a-ye-nvhst-ayvtho:-ko. Akwe: tsi t-ka-nvhst-ayvth-u yvyakwe' tanu y-v-yak- wa-hroht-v-ht-e. E-tho ne ó:nv v-yak-wa-nor-oht-hsi-'. → then there.it.set there-would-one- corn-plant-REVERS all to there-it-corn-planted will.we.go and there-will-we-all-ear-fall- CAUSE-PUNC then the then will-we-all-husk-stand-REVERS-PUNC | |||
“Then it was time to harvest it (the corn). We would all go to the cornfield and take it (the corn) from the stalks. We would then husk it.” | |||
While Minhast uses NI for backgrounding purposes as in Mohawk, its motivation for backgrounding an entity through NI is significantly different from Mohawk and the rest of the Iroquioan languages. Minhast employs NI to keep the IN available so that it can later restore it to core status, thereby returning it to the foreground in the discourse. Thus, in Minhast the status of an IN is somewhere between an oblique argument and a core argument, a key difference between how it uses NI as opposed to Mohawk. The application of NI in Minhast serves as a foreshadowing technique; temporarily backgrounded, the IN often resurfaces later in the discourse as a principal player (core argument) of the narrative. The foreshadowing function of NI also demonstrates an important point: INs occurring in | |||
21 | |||
extended discourse are highly referential in Mohawk, as argued by Baker (1996). Minhast, based on native speakers’ self-reports. | |||
The same is true in | |||
Additionally, there is a type of Control clause, namely the Causative clause22 which requires the | |||
21 This point seems to be contested by others, at least in the case of Mohawk. See Mithun 2010:12. | |||
22 In addition to Causative clauses, Control clauses also include Cognitive clauses, Desiderative clauses, Mandative | |||
clauses, Decision clauses etc. Causative clauses differ from other types of Control clauses in that Causative clauses are also Previous Event-Successive Event conjoined clauses (PEC-SEC), always recognizable by the Subordinator affix - | |||
combination of NI and Antipassivation if the speaker wishes to reduce valency. As with all Control clauses, Causative clauses are highly transitive and increase valency. As in all Control clauses, they are recognizable by the presence of two subjects, one called the Controller, which manipulates or acts on a secondary subject, called the Controllee. The Controllee is responsible for carrying out the actual event or achieving the final end state. | |||
Controllers are typically Ergative arguments, and Controllees are cast to Absolutive status, as Sentence 13a demonstrates: | |||
13a) Yakte anxekte išpisattabeddadekarumā siyyekittūmarraran . | |||
Yakte šp-sattabe-redad-ek-ar-u=mā siyyekit-tūman-ar-an | |||
1S.ERG brother.ABS CAUS-kill-man-1S.ERG+3S.ABS-PST-TRANS=SUBORD burn-house- PST-INTRANS) | |||
“I sent my brother to kill the man and then he (i.e. the brother) burned down the house.” | |||
Sentence 13a targets the brother as the Controllee, and so was cast as an Absolutive. However, sometimes valence-reducing operations are required to convert or retain the Controller as the S/O pivot. This can be done only by demoting or removing the Controllee23 acting on the semantic Patient. In such a case, the logical PT is incorporated into the verb in the matrix. The Controllee can now be removed from the Absolutive position, leaving it vacant for the Controller to move into the vacated Absolutive space. The combination of NI and Antipassivation yields the example in Sentence 11b: | |||
13b) Yak išpisattabeddadekarampamā siyyekkittūmanekarampi. | |||
Yak šp-sattabe-redad-ek-ar-an-pi=mā [PRO] siyyekkit-tūman-ek-ar-an. | |||
S.ABS CAUS-kill-man-1S.ABS-PST-INTRANS=SUBORD burn-house-1S.ABS-INTRANS “I sent someone to kill the man, and then I burned down his house.” | |||
The Absolutive argument, yak, is now available to feed the S/O pivot, thanks to the coordination of both NI and Antipassivation to reduce the valency of the sentence; NI backgrounds the logical PT redad, and the application of Antipassivation allows yak to occupy the Absolutive position. Yak can thus serve as the Absolutive feeder of an S/O pivot. | |||
Other Control clauses differ from Causative clauses in that they are embedded in their matrix clauses as | |||
24 | |||
complements, i.e. nominalizations marked with the clitic =naft . Cognitive and Perception clauses are | |||
mā joining the PEC to the SEC. | |||
23 The Causative Controllee can be brought back into the clause, but it must appear as an Instrumental oblique argument. | |||
The reading could then be calqued: “By using [NP], I caused to kill the man so that I may live.” Languages that have the Passive voice, such as English, often use an Instrumental oblique marker on a demoted subject just as languages with the Antipassive do with Instrumental oblique arguments. Compare English “I was caused by Joe to kill the man so that I could live” with Minhast “Yak anxē=par išpisattabeddadekarampamā siyyekkittūmanekaran”, lit. “By using my brother, I caused to kill the man and then I burned down the house”, where the Controllee, anxea (“brother”), is retained as a demoted subject via Antipassivation. | |||
24 The use of nominalizations to serve as complements can lead to ambiguities: Sap min Aynuwast gubbāturrād sattabeddadaranaft yakte sarekaru can be interpreted as either “I saw this Ainu soldier who killed the man” or “I saw | |||
included in this subtype: | |||
14a) Sap min Aynuwast min gubbāturrād sattaberedadaranaft yakde nusilleku | |||
25 | |||
Sap min Aynuw-ast min gubbāturrād sattabe-redad-ar-an=naft yak=de nusill-ek-u. | |||
DEM.PROX CONN aynu.person-GENT CONN warrior kill-man-PST-INTRANS=NMLZ | |||
1S=ERG know.a.fact-3S.ABS + 1S.ERG-TRANS | |||
“I know that this Ainu warrior killed the man” or “I know this Ainu warrior who killed the man.” | |||
14a.1) Sattabe-redad-ar-an=naft, yak=de sap min Aynuw-ast min gubbāturrād nusill-ek-u. | |||
“ | |||
14b) | |||
u. | |||
14c) Yak nusill-ek-an-pi. “I know a fact [about something]”. | |||
14d) Yak nusill-redad-ek-an-pi. “I know a fact [regarding something about] the man.” | |||
14e) Sap min Aynuwast min gubbāturrād sattabe-redad-ar-an=naft yak=de sar-kaddara-ek-ar- u. “I event-saw him, this Ainu warrior killed the man.” | |||
14e.1) Sattabe-redad-ar-an=naft, yak=de sap min Aynuwast min gubbāturrād sar-kaddara-ek- ar-u. “I event-saw this Ainu warrior, he that killed the man.” | |||
14f) Redad=aran sap min Aynuwast min gubbāturrād sattabe-ar-an-pi=naft yak=de sar- kaddara-ek-ar-u. “I event-saw him, this Ainu warrior killed a man” or “I event-saw him, this Ainu warrior who killed a man”. | |||
14f.1) Redad=aran sattabe-ar-an-pi=naft yak=de sap min Aynuwast min gubbāturrād sar- kaddara-ek-ar-u. “I event-saw this Ainu soldier, he that killed a man”. | |||
14g) Yak sar-ek-ar-an-pi. “I saw [something].” | |||
14h) | |||
that this Ainu soldier killed the man.” With non-third person pronouns this ambiguity disappears: Sattabeddadtaharanaft yakte sartahekaru (lit: “I saw you that killed the man”) can only be interpreted as a complement. Another strategy for disambiguation is NI of -kaddar(a)- “event” or -razn- “matter, issue, subject” to the Control verb, e.g Sap min Aynuwast gubbāturrād sattabeddadaranaft yakte sar-kaddara-ekaru (“lit: “I event-saw the Ainu who/that killed the man.”). However, NI of -kaddar(a)- and -razn- are optional, and more often than not omitted. | |||
. | |||
kill-man-PST-INTRANS=NMLZ 1S=ERG DEM.PROX CONN aynu.person-GENT | |||
know.a.fact-3S.ABS + 1S.ERG-TRANS | |||
He who killed the man, I know this Aynu warrior (did it)” or “I know this Ainu warrior who | |||
killed the man.” | |||
Redad=aran sap min Aynuw-ast min gubbāturrād sattabe-ar-an-pi=naft yak=de nusill-ek- | |||
“I know that this Ainu warrior killed a man” or “I know this Ainu warrior who killed a man.” | |||
25 The verb nusill- alleviates the ambiguities mentioned in footnote #17, since its full meaning is “to know a fact”. | |||
a) ** Yak sar-redad-ek-ar-an-pi. “I man-saw.” Antipassivation + NI is not allowed here. | |||
b) Anxea yak sar-redad-ek-ar-an-pi. “I man-saw (my) brother.” Antipassivation + NI is allowed here because this is Type IV NI (see below) | |||
14i) Sap min Aynuwast min gubbāturrād sattabe-redad-ar-an=š=naft yak=de ittawas-ek-ar-u. 14j) Redad=aran sap min Aynuwast min gubbāturrād sattabe-ar-an-pi=naft yak=de ittawas-ek- u. | |||
14k) Yak ittawas-ek-ar-an-pi. | |||
14l) **Yak ittawas-redad-ek-ar-an-pi. | |||
Notice that if the Controller is placed into the Absolutive position by Antipassivation (Sentences 14d, 14h, 14l) the nominalization must be deleted, since Minhast does not tolerate two separate Absolutive arguments and nominalizations must always have an Absolutive argument to relativize on. The IN of the complement clause in which the Controllee is embedded in may be retained by incorporation into the Control verb, but without nominalization, there is no way to qualify the NI. That must be determined by context. | |||
=== Type IV Noun Incorporation - Classificatory Functions === | |||
== | Some languages have expanded the range of functions that NI can perform beyond valence operations and discourse pragmatics. These languages take NI to an advanced level, whereby a generalized noun is incorporated to classify or categorize the Patient, which has more specific meaning. Again, Mohawk is an exemplary language that exploits this form of NI pervasively, called Classificatory NI (Mithun's Type IV category). The following passage from Mithun (1984) illustrates this NI function type: | ||
== | 14a) Tohka niyohserd:ke tsi nahe' sha'te:ku niku':ti rabahbot wahu-tsy-ahni:nu ki rake'niha. → several so.it.year.numbers so it.goes eight of.them bullhead he-fish-bought this my.father. | ||
== | “Several years ago, my father bought eight bullheads.” | ||
== | In this example, the incorporated noun (i)tsy (“fish”) is co-referrent with the Patient rabahbot (“bullhead fish”); (i)tsy serves as a general modifier to classify the more specific rabahbot. As the narration continues, rabahbot is subsequently referred to by the incorporated noun (i)tsy: | ||
14b) Saháhkete' ki:kv rakenuhá:'a s-a-h-vtsy-ahsheruny-à:na-'. Yusa :rawe ki': óksa'k wa-h-vtsy- ahserul:ni tanu wa-h-vtsy-akeri:tahw-e. Tsi n-a-ho-tsy-ari-hs-e ki' ki:kv wahv:ru, “Tho yukyatv:ro rinu-tsy-anut-v-:ra.” | |||
back.he.turned this my.uncle back-PAST-he-fish-fix-go.to-PUNC back-he-arrived just quick PAST-he-fish-fix(PUNC) and PAST-he-fish-fry-PUNC as as-PAST-him-fish-fry-finish-PUNC just this he.said there we.two.friends.are I/him-fish-feed-for-go.to | |||
“My uncle then returned to fix them (the fish). At home, he cleaned and fried them (the fish), and when they (the fish) were ready, he decided to take them (the fish) over to his friend as a special treat.” | |||
The Mohawk examples show how the generic IN, (i)tsy, is used like an agreement marker throughout the entire discourse. Although it is generic, it is used to refer back repeatedly to the more specific rabahbot (bullheads), which is mentioned only once, at the beginning of the narrative. | |||
Minhast, like Mohawk, uses Classificatory NI, but once again, the factors motivating Minhast NI to | |||
implement Classificatory NI are different. As mentioned earlier, certain noun classes are resistant to NI or forbidden. Nouns that are resistant to NI or forbid it are typically located in the high end of the animacy scale. Nouns that fall under this portion of the animacy hierarchy are kinship nouns and proper nouns. | |||
As in Mohawk, for Classificatory NI Minhast incorporates a generic noun. This incorporated argument is then used as an agreement marker to the NI-resistant noun, which is always in the Absolutive. A requirement for Classificatory NI in Minhast is that the noun which the IN is corefererent with cannot be a semantic Agent, and for this reason, once again Antipassivation is blocked by NI. | |||
The NI-resistant noun and the generic IN that serves as its agreement marker are highlighted in bold underlined font: | |||
15a) Supnar min anxekte Keyyis min niħkašektaran duntittarraru duntittarumā, anxēa indintanuskikirimredadnnarumā. | |||
Supnar min anxē=ek=de Keyyis min niħkaš-ek=de=aran dut-nittam-ar-u=mā, anxēa nd-inta- nusk-kirim-redad-nn-ar-u=mā... (PROP.NOUN CONN brother=3MS.ABS+1S.ERG=ERG PROP.NOUN CONN friend-3S.ABS+1S.ERG=ERG=DAT DAT.APPL-shout-PST- TRANS=SUBORD, brother.ABS INCEP-INTENS-ADVER.APPL-speak-man- 1P.EXCL.ERG-PST-TRANS=SUBORD) | |||
“My brother Supnar shouted at my friend Keyyis and we started to argue against my brother...” | |||
The narrative continues. The narrator decides to cast himself as the Absolutive argument to feed the S/ O pivots in the succeeding clauses. When he needs to mention his brother as the Patient, he chooses not to use the Antipassive in order to cast his brother as a Dative oblique; to do so would imply the brother is incidental information, which is not the case in this passage. Instead, he applies NI on redad, which coindexes his brother as the logical Patient. Redad then serves as a proxy for his brother for NI operations: | |||
15b) ...kūdāš segwekarammā, (kūde) yummatekarumā, rabbaddadekarammā, karyaħtendepār kaħmadekarampamā šarrataran ušniddadekarammā, nittarredadekarannamā: “Bakran wattaħte ušnktahuš? Hatā anxēšattarakš? Ta'astakkemarunaft wastānešattarakte hittastānehakkemaruš?” | |||
...kū=dāš segw-ek-ar-an=mā, (kua=de) yummat-ek-ar-u=mā, rabba-redad-ek-ar-an=mā, karyaħt-enn=de=pār kaħmad-ek-ar-an-pi=mā šarrat=aran ušn-redad-ek-ar-an=mā, nittam- redad-ek-ar-an=namā: “Bakran wa=tah=de ušn-ktah-u=š? Hatā anxēa-šattar-hak=š? Ta=ast-hakkem-ar-u=naft wastāne-šattar-hak=de hitt-wastāne-hakkem-ar-u=š?” | |||
3S.OBL=MAL make.fist-1S.ABS-PST-INTRANS=SUBORD, (3S=ERG) swing.fist- 1S.ABS+3S.ERG-PST-TRANS=SUBORD, grab-man-1S.ABS-PST-INTRANS=SUBORD, arm-3S.NEUT.ANIM.ABS+3S.ERG=ERG=INSTR twist-1S.ABS-PST-INTRANS- ANTI=SUBORD earth=DAT hit-man-1S.ABS-PST-INTRANS =SUBORD, shout.at-man- 1S.ABS-PST-INTRANS=DIR.QUOT: why CONN=2S=ERG hit-1S.ABS+2S.ERG- TRANS=IRREAL? NEG brother-RECIP-1P.INCL.ABS=IRREAL? NEG= beget- 1P.INCL.ABS+3P.COMMON.ERG-PST-TRANS=NMLZ blood-RECIP - 3S.NEUT.ANIM.ABS+1P.INCL.ERG=ERG give-blood-1P.INCL.ABS+3P.COMMON.ERG- | |||
PST-TRANS=IRREAL | |||
“...[and] we started to fight, my brother and I. He swung at me (but missed), and I grabbed my brother, then I twisted his arm and struck him to the ground. I yelled at my brother, 'Why do you strike at me? Am I not your brother? Are we not of the same blood? (lit. Our mutually- shared blood, did not those that begat us give us blood?)'” | |||
While these passages demonstrate that Minhast does employ Classificatory NI, it does not do so as extensively as in Mohawk. In Minhast, Classificatory NI is employed to get around the obstacles presented by NI-resistant nouns, which by their nature, rank high in the animacy hierarchy. Thus, Classificatory NI is rarely, if ever, encountered in passages with low-animacy entities. | |||
Most important, though, is whether Classificatory NI is required to maintain the S/O pivot. If not, Classificatory NI is not employed. Thus, one is not likely to see a sentence like in 16a. In fact, a native speaker would find it quite odd, and most likely ungrammatical: | |||
16a) Dūy zaydakkī sarekaru, (dūy) rimar-sankūy-ekarumā, (dūy) niyyet-sankuy-ekaru. | |||
“I saw the salmon in the river, I fish-speared it, then I fish-pulled it out of the water.” | |||
The sentence does not require Classificatory NI because dūy (“salmon”) is incorporatable and no advantage is gained by using the generic noun sankūy (“fish”) to co-index dūy. While both nouns are animate, they are also neuter in gender, which when compared to nouns that have masculine and feminine nouns, both of these nouns lie lower in the animacy spectrum; Classificatory NI is typically employed with highly animate nouns, which include proper nouns and kinship terms. However, dūy is low in the animacy scale, so the incorporation of the more generic term sankūy for Classificatory NI is unnecessary. | |||
Compare this with the previous Mohawk example of Classificatory NI: | |||
16a) Tohka niyohserd:ke tsi nahe' sha'te:ku niku':ti rabahbot wahu-tsy-ahni:nu ki rake'niha. several so.it.year.numbers so it.goes eight of.them bullhead he-fish-bought this my.father. | |||
Here rabahbot (the bullhead fish) and generic, incorporated noun tsy (fish) lie in the same level on the animacy hierarchy, but Mohawk utilizes Classificatory NI so that tsy can function as an agreement marker for rabahbot later discourse. | |||
In the case of Minhast, the application of Classificatory NI background dūy conveys no observable benefits, as both dūy and sankūy are at the same animacy level. Minhast uses Classificatory NI when the target noun is unincorporatable, as in the case of proper nouns and kinship nouns. However, neither dūy nor sankūy are considered ineligible for incorporation. | |||
If dūy is the topic of interest, it must be cast as the Pivot, which requires that it assume the role as an Absolutive argument: | |||
16b) Dūy zaydakkī sarekaru, [PRO] rimarekarumā, [PRO] niyyetekaru. | |||
“I saw the salmon in the river, speared it, and pulled it out of the water.” | |||
Or if the pronoun yak/-ek- were the topic of interest, it would be the S/O pivot, and then dūy can become the incorporated Patient: | |||
16c) Dūy=aran zaydakkī sarekarampi, rimar-dūy-ekammā, [PRO] niyyet-dūy-ekan. | |||
“I saw some salmon in the river, I salmon-speared, then salmon-pulled-out-of-the-water. | |||
...or dūy can be nominalized, in which it would again serve as the S/O pivot: 16d) Dūy zaydakkī sarekarunaft rimarekarumā [PRO] niyyetekaru. | |||
“The salmon that I saw, I speared and pulled it out of the water.” | |||
However, the incorporation of dūy would make it ineligible to function as the Pivot. The following | |||
** Sartuyekarammā rimarekaru >> ** sar-duy-ek-ar-an-mā rimar-ek-ar-u | |||
** “I fish-saw then speared it.” | |||
Here the intent of the speaker was to mark duy as a PT in the first clause via NI, then link the first | |||
clause to the second clause via the -mā suffix. The verb in the first clause is intransitive after the | |||
incorporation process, which is why it received intransitive marking with the -an- affix. The second | |||
clause is marked as a transitive verb by the suffix -u. The speaker used the transitive -u suffix thinking | |||
that its PT argument was the incorporated -duy-, however this is ungrammatical. The transitive marker | |||
requires an Absolutive argument, but the incorporation of -duy- removed it from the Absolutive | |||
position, thus there no longer is an Absolutive argument available to rimarekaru to function as the | |||
Pivot. In Minhast, an IN can never serve as a Pivot in clause chains, only an explicit NP in the | |||
Absolutive case frame, or null-marked implicit NP, whose agreement marker is the ABS agreement | |||
marker inside the verb complex. To make this sentence grammatical while retaining the IN, at the | |||
minimum an Applicative affix marking an explicit or implicit NP that has been promoted to core status | |||
26 | |||
The fact that animacy conditions have to be met before Classificatory NI is applied explains a major reason why it is not as extensively used in Minhast as it is in Mohawk. The requirements of the S/O pivot, which is exploited pervasively in Minhast, also has a role in governing whether Classificatory NI is employed. Classificatory NI appears to be a last-resort measure that is used if no other valence operation can feed the S/O pivot with the proper argument; otherwise, it is not employed. That Minhast has other syntactic mechanisms, such as the valence operations of Antipassivation and Applicative Formation, as well as nominalizations and the wa-clause connective and allied structures all contribute to the low frequency of Classificatory NI in Minhast. | |||
The situation for Mohawk is the inverse of the Minhast system: lacking the ancillary syntactic mechanisms for altering argument structure found in Minhast, Mohawk needs Classificatory NI because the language does not employ nominalizations. Mohawk clauses, even when chained together into long clause chains, are still ultimately regarded as truly independent, stand-alone sentences; stripping one of the component clauses from its matrix clause does not change its grammaticality. In | |||
26 The 3S.ABS+1S.ERG agreement marker is -ek-, which is homophonous with 1S.ABS -ek-. The Transitive verb suffix - u disambiguates which -ek- form is being used. | |||
is required. | |||
contrast, nominalizations cannot stand alone, they must be contained within a matrix clause; stripping them from their matrix clause does make them ungrammatical. Minhast has its S/O pivot system available to it, which it exploits aggressively, especially in the case of nominalizations; Mohawk has no such syntactic construct. That Mohawk has the Classificatory NI available to it to handle the narratives in 14a and 14b does not make it superior to Minhast. That Minhast can use nominalizations and its S/O pivot system to handle the same narrative and minimize using Classificatory NI does not make it superior to Mohawk. They are simply different syntactic structures available to their respective languages, and each language maximizes the tools available to solve syntactic problems such as the ones just described. | |||
* This is an example of Mithun's Classificatory NI (Class IV) being exploited by Minhast. Here, it is essentially creating the equivalent of a locative noun in other languages, e.g. English "Within the ''interior of'' the beast..." Here, ''nua'' means "side" has been incorporated into the verb complex. The implicit head is ''suharak'' (deerskin), which was mentioned in a previous line in the passage, referred to by the Locative applicative ''naħk-''. This construction is equivalent to saying "Next to it", "By its side", etc. | * This is an example of Mithun's Classificatory NI (Class IV) being exploited by Minhast. Here, it is essentially creating the equivalent of a locative noun in other languages, e.g. English "Within the ''interior of'' the beast..." Here, ''nua'' means "side" has been incorporated into the verb complex. The implicit head is ''suharak'' (deerskin), which was mentioned in a previous line in the passage, referred to by the Locative applicative ''naħk-''. This construction is equivalent to saying "Next to it", "By its side", etc. | ||
Line 18: | Line 263: | ||
| translation = Canonballs explode next to it the deerskin. | | translation = Canonballs explode next to it the deerskin. | ||
}} | }} | ||
= Truncation/Weak Suppletion = | = Truncation/Weak Suppletion = | ||
Line 44: | Line 284: | ||
Similarly, Minhast INs exhibit weak suppletion, and it occurs extensively, particularly with nouns longer than two syllables, e.g. ''sussagarānī'' > ''-suggan-'' ("big toe"). | Similarly, Minhast INs exhibit weak suppletion, and it occurs extensively, particularly with nouns longer than two syllables, e.g. ''sussagarānī'' > ''-suggan-'' ("big toe"). The contrast can be seen in the following two examples, the first where the noun occurs in its full form as the dependent argument in a possessive NP, and the second wherein the noun appears in truncated form after noun incorporation: | ||
{{Gloss | {{Gloss | ||
|phrase = | |phrase = <u>Sussagarānī</u>tirektiki kahušnišattekaran. | ||
| IPA = | | IPA = | ||
| morphemes = sussagarānī-tirek=de=ki kah-ušn-šatt-ek-ar-an | | morphemes = sussagarānī-tirek=de=ki kah-ušn-šatt-ek-ar-an | ||
| gloss = big.toe-3SN.INAN.POSSM+1S.POSSR=ERG=LOC INV.VOL-hit-RFLX-1S.NOM-PST-TRNS | | gloss = big.toe-3SN.INAN.POSSM+1S.POSSR=ERG=LOC INV.VOL-hit-RFLX-1S.NOM-PST-TRNS | ||
| translation = I | | translation = I stubbed my big toe. | ||
}} | }} | ||
{{Gloss | {{Gloss | ||
|phrase = | |phrase = Kahušni<u>sugga</u>š<u>n</u>attekaran. | ||
| IPA = | | IPA = | ||
| morphemes = kah-ušn- | | morphemes = kah-ušn-sussagarānī-šatt-ek-ar-an | ||
| gloss = INV.VOL-hit-toe-RFLX-1S.NOM-PST-TRNS | | gloss = INV.VOL-hit-toe-RFLX-1S.NOM-PST-TRNS | ||
| translation = I stubbed my big toe. | | translation = I stubbed my big toe. | ||
Line 65: | Line 305: | ||
The pattern of truncation is unpredictable; syllable loss may occur in initial, medial, or final positions, although noun roots with more than two syllables tend to lose either their medial or final syllables and retain the initial syllable, but exceptions abound, such as ''allāga'' > ''-lgagg-'' (conch) . | The pattern of truncation is unpredictable; syllable loss may occur in initial, medial, or final positions, although noun roots with more than two syllables tend to lose either their medial or final syllables and retain the initial syllable, but exceptions abound, such as ''allāga'' > ''-lgagg-'' (conch) . | ||
= Noun Incorporation of Oblique Arguments= | |||
In many languages which exhibit noun incorporation, the type of noun that can be incorporated into the verb is often restricted. Some languages incorporate body parts only, others are restricted to inalienable nouns or some other semantic category. Other languages that exhibit extensive noun incorporation, of which Mohawk and its relatives in the Iroquoian language family are the most studied, while having much fewer semantic restrictions, still limit the syntactic or thematic role of the noun that can be incorporated: these are that of the Patient argument, and in some cases the Instrument argument. Other arguments serving in a different thematic/theta role are barred from incorporation. Other noun incorporating languages, such as Chukchi, appear to have no restrictions on the theta role of the incorporated noun (IN); but when these oblique nouns are incorporated, the only way to recover their thematic role is by context alone. | |||
Minhast is one of those languages that can incorporate oblique arguments. However, the oblique arguments that can be incorporated are constrained by the semantic characteristics of the verb. Some transitive verbs which require a third argument, which is always an oblique noun, can optionally incorporate the oblique noun. The verb wasaskiyu is such an example: | |||
wasaskiyu - “to put something on an object” (E.g. “Please put it on the chair”) | |||
Verb takes 3 arguments, v(Agt, Pt, Obl.LOC) | |||
Agent is typically ERG | |||
Pt is typically Abs | |||
Obl.LOC can be incorporated. | |||
Surma dūy wasaskizekyašennaru. | |||
Surma dūy wasaski-zekyaš-enn-ar-u | |||
PN salmon put-table-3S.ANIM.ACC+3MS.ERG-PST-TRN | |||
“He put the salmon on the table.” | |||
If this were Ainu, the incorporated Locative noun would license an Applicative affix; in Mohawk (and presumably its closely related Northern Iroquoian relatives), I haven’t seen incorporation of an oblique, only Patients. An exception is found in Tuscarora (Mithun p. 201 “Word Orders”, example 12a, contrasting non-incorporated 12b) in which we see a Locative arg incorporated in the stative verb | |||
Chukchi incorporates obliques, but often the theta role must be inferred by context. See A Lexical Account of Noun Incorporation in Chukchi (Silke Lambert, p.56). Minhast, as demonstrated in this article, incorporates obliques like Chukchi, but is much more restricted based on the semantic characteristics of the verb, i.e. verbs that incorporate obliques are restricted to a limited set of theta roles, and often the number of roles is restricted to one. | |||
The semantic properties of a verb, namely the type of semantic/thematic role of an NP it can accept as an argument, influence what nouns may be incorporated. This semantic property limits the scope of which oblique nouns can incorporate, a limitation not exhibited in Chukchi. Motion verbs typically incorporate Goal or Origin nouns; Positional verbs typically incorporate Locative nouns. Otherwise, transitive verbs may incorporate either Patient or Instrument nouns; this sort of incorporation may seem to create ambiguities, but such is not the case, as both the polypersonal agreement pronominal affixes, and the presence or absence of an Instrumental Applicative make clear whether the Absolutive is a Patient or is an Instrument that has been promoted to the Absolutive argument. A few examples: | |||
Example of a Goal argument: | |||
Iknatumankaran. | |||
Ikna-tūman-ek-ar-an | |||
go-house-1S.NOM-PST-INTR | |||
“I went to/towards the house.” | |||
Example of an Origin argument: | |||
Hahurtaħran. | |||
ha-hūr-tah-ar-an | |||
come-mountain-2S.NOM-PST-INTR | |||
“You came from the mountain.” | |||
Example of a Locative argument: | |||
Sap puħtabanakkaran. | |||
Sap puħta-banak-ek-ar-an | |||
This stand-rock-1S.NOM-PST-INTR | |||
“I stood on this rock.” | |||
Example of a Patient ABS with incorporated Instrumental argument: | |||
Redadesap ušnišuhapnekarun. | |||
Redad=sap ušn-šuhapna-ek-ar-un | |||
Man this strike-sword-3MS.ACC+1S.NOM-PST-TRNS | |||
“I struck the man with this sword.” | |||
Example of an Instrument ABS with incorporated Patient argument: | |||
Sapim šuhapna matušnerdattirkarun. | |||
Sap min šuhapna mat-ušn-redad-tirk-ar-un | |||
This CONN sword INST.APPL-strike-3NS.ACC+1S.NOM-PST-TRNS | |||
“With this sword I struck the man.” | |||
Noun incorporation is often associated with clauses that are structurally transitive, regardless of whether or not an oblique argument has been promoted to a core argument. However, in Minhast some stative verbs can noun incorporate. In other words, under certain circumstances, a clause that is structurally intransitive may also undergo noun incorporation. Such intransitive clauses tend to be nouns whose single core argument’s theta-role is that of Experiencer, e.g. | |||
6) Kuldantuhamaran. | |||
kuldan-tuham-ar-an | |||
sick-fever-3S.NOM.PST-INTR | |||
“He is sick with fever/He is sick and feverish.” | |||
7) šuhapna wastanxundēban. | |||
šuhapna wastan-xunde-ab-an | |||
Sword bleed-wound-3S.NOM.IMPF-INTR | |||
“He is bleeding from his sword-wounds/He is bleeding and wounded by sword/Because of that sword he was bleeding and wounded.” | |||
8) Saxtisuspaħtayattaran. | |||
saxt-suspaħ-tayatta-ar-an | |||
INCH-be.blind-poison-3S.NOM.PST-INTR | |||
“He became blind because of the poison.” | |||
As an additional observation, those stative verbs that can incorporate tend to indicate sickness, injury, or congenital or other physical defects (as in Ex. #8 above). When the Inchoative prefix -saxt- occurs with the incorporated noun, native speakers tend to indicate the IN is the direct cause of the Experiencer’s state (Ex. #8). Otherwise the IN provides further details of or delimits the Experiencer’s current state (Examples #6 & #7), hence the alternative translations using the conjunction “and”. But this is not always so, as illustrated in the case of Ex. #7 which lacks the Inchoative. Here the sentence without -saxt- would seem to suggest that something else may have caused the bleeding, but there were other wounds that resulted from a sword. However, context and general knowledge that swords tend to cause bleeding would make that interpretation dubious. The use of the Inchoative -saxt- would definitely dispel that ambiguity, but is not necessary if context is sufficient to disambiguate between the two possible interpretations. | |||
Also note in Ex. #7 the stranded NP “šuhapna” (sword): this stranded NP is the modifier of the incorporated noun “xunde”; the equivalent non-noun incorporated sentence would be “šuhapna min xundeyār wastanaban” (lit: From sword-wounds, he was bleeding), where the Ablative clitic =yār indicates the cause of the subject’s bleeding. The polypersonal agreement affix in Ex. #7 indicates there is only one core argument only, which would be the logical subject. This clearly demonstrates that “šuhapna” is a stranded NP. | |||
= Noun Incorporation in Intransitive Verbs = | = Noun Incorporation in Intransitive Verbs = | ||
Line 174: | Line 488: | ||
| IPA = | | IPA = | ||
| morphemes = purrak saxt-raħk-tahal-ruppamak-ek-ar-an | | morphemes = purrak saxt-raħk-tahal-ruppamak-ek-ar-an | ||
| gloss = pigment=ABL | | gloss = pigment=ABS INCH-APPL.ABL-be.green-face-1S.NOM-PST-INTR | ||
| translation = My face became green from the dye (lit. "Because of the pigment, I became green-faced.") | | translation = My face became green from the dye (lit. "Because of the pigment, I became green-faced.") | ||
}} | }} | ||
Line 189: | Line 503: | ||
}} | }} | ||
= Valency and Agreement Marking Irregularities = | |||
It remains debatable as to whether stative verbs with incorporated meteorological nouns are monovalent or zero-valent. The following example lends support to a monovalent interpretation: an overt non-null pronominal agreement marker ''-i-'' indicates that an underlying third person inanimate plural absolutive argument exists and has undergone pro-drop: | It remains debatable as to whether stative verbs with incorporated meteorological nouns are monovalent or zero-valent. The following example lends support to a monovalent interpretation: an overt non-null pronominal agreement marker ''-i-'' indicates that an underlying third person inanimate plural absolutive argument exists and has undergone pro-drop: | ||
Line 200: | Line 514: | ||
}} | }} | ||
= Polypersonal Marking = | |||
The primary purpose of the polypersonal markers in the Minhast verb are to reference the core arguments of its clause, whether they appear overtly, or are omitted through pro-drop. However, polypersonal marking can target the IN under certain restrictions. When agreement marking does appear, they tend to occur with collective or mass nouns, such as the inherently collective ''iyuššit''. Moreover, this type of incorporation occurs with certain verbs only, particularly ones indicating movement, and even then ''iyuššit'' triggers agreement marking in a small fraction among these verbs. | |||
{{Gloss | |||
|phrase = Nayyakiyuššitiyaran. | |||
| IPA = | |||
| morphemes = nayyaki-iyuššit-i-ar-an | |||
| gloss = gather.together-storm.cloud-3.ANIM.P-PST-INTR | |||
| translation = Storm clouds gathered. | |||
}} | |||
Oftentimes irregularities in gender-number concord may appear. An example where agreement marking is lacking with the very same collective noun follows in the next example. Note that the verb ''-adu-'' ("be many"), unlike ''-nayyaki-'', is not a motion verb. The lack of an overt agreement marker with ''-adu-'' points towards a zero-valent interpretation: | |||
{{Gloss | {{Gloss | ||
Line 220: | Line 545: | ||
}} | }} | ||
Interestingly, ''-puht-'' can license agreement with other nouns, such as ''kayyūn'' "tree", when a collective meaning is intended: | |||
{{Gloss | |||
|phrase = Yaššapuħtakayyummaharan. | |||
| IPA = | |||
| morphemes = yašša-puħt-kayyūn-mah-ar-an | |||
| gloss = there.DIST-stand.upright-tree-3S.NEUT-PST-INTR | |||
| translation = The trees stood there. | |||
}} | |||
Diachronic factors may explain the irregularities involving agreement marking for a subset of incorporated nouns interacting with a subset of verbs. The Proto-Nahenic ancestor originally had an extensive hierarchical noun class system, remnants of which remain in Minhast's relative Nahónda as evidenced by even more irregularities in the latter, and in its other relative Nankôre, whose elaborate nominal hierarchy may be a preservation of the protolanguage's original noun class system or an extensive elaboration of it. The irregular agreement marking triggered by ''-iyuššit-'' among a subset of a select class of verbs suggests that the noun once fell within a noun class of a particular animacy level. When the protolanguage split, the original noun class system were restructured in the daughter languages; further reductions and loss, particularly in both Minhast and Nahónda, left a residue in the form of the irregular agreement marking seen today. | Diachronic factors may explain the irregularities involving agreement marking for a subset of incorporated nouns interacting with a subset of verbs. The Proto-Nahenic ancestor originally had an extensive hierarchical noun class system, remnants of which remain in Minhast's relative Nahónda as evidenced by even more irregularities in the latter, and in its other relative Nankôre, whose elaborate nominal hierarchy may be a preservation of the protolanguage's original noun class system or an extensive elaboration of it. The irregular agreement marking triggered by ''-iyuššit-'' among a subset of a select class of verbs suggests that the noun once fell within a noun class of a particular animacy level. When the protolanguage split, the original noun class system were restructured in the daughter languages; further reductions and loss, particularly in both Minhast and Nahónda, left a residue in the form of the irregular agreement marking seen today. |
Latest revision as of 20:14, 26 October 2024
This article is a construction site. This project is currently undergoing significant construction and/or revamp. By all means, take a look around, thank you. |
Introduction
Types of Noun Incorporation
Mithun's Framework
Languages which employ NI do have alternative, analytic structures containing the same semantic information as an NI structure does. However, NI provides a way to manipulate discourse, reduce the salience of an entity in order that other entities can take precedence in extended speech, provide stylistic and rhetorical alternatives to their corresponding analytic expressions, and even derive new lexical items. Mithun (1984) has identified four categories of NI that occur cross-linguistically. The four categories Mithun has identified are:
- Type I - Lexical Compounding: the creation of new lexical items by compounding a noun root and verb root;
- Type II - Case Manipulation: a noun (usually a Patient, although Instrumental and Locative nouns may be involved) is incorporated into the verb complex. This is a valence operation: if the incorporated noun was originally a core argument, another argument can occupy the position vacated by the IN and assume core status. Alternatively, depending on the semantic nature of the verb, Oblique8 nouns that are Instruments, Locatives, or Goals may also be incorporated;9
- Type III - Manipulation of Discourse: NI is used to background information in sections of the discourse so that other arguments are brought to the foreground. It allows speech participants to focus on the important entities within a particular passage of the discourse;
- Type IV - Classificatory NI: Mithun describes this type of NI wherein a “...relatively general N(oun) stem is incorporated to narrow the scope of the V(erb)...but the compound noun stem can be accompanied by a more specific external NP which identifies the argument implied by the IN.”
Not all languages that employ NI exhibit all four types. For example, in the indigenous Brazilian language Karajá, NI is mostly restricted inalienably possessed nouns, particularly body parts, and does not alter valence, which by nature involves Type II case manipulation (Ribeiro 2001). Chukchi freely employs Type II NI to alter argument structure, Type III is attested per Mithun, but Type IV NI is absent (Modena & Muro). Nevertheless, languages that employ all four types of NI are found in disparate languages, including Mohawk, Caddo, and Gunwinggu. Minhast also falls within this group.
Type I Noun Incorporation - Lexical Compound Formation
Minhast utilizes Type I NI to create lexical compounds but only if a particular activity, state, or event occurs frequently to warrant institutionalizing into the lexicon. Usually, one or both elements of the compound are shortened, as in the following examples:
5a) aydann- “To store water in a cistern, reservoir, or catch-basin” (derived from ayaya - “put something into a container” + dannua “water”)
5b) uzdann- → uzzat-dannua “To draw water from a well” (derived from uzzat- “to pull
something out of a container or other enclosing object, e.g. an envelope” + dannua “water”
Compounding NI is a derivational process. If the compound yields a new verb, it has the fullstatus and capabilities of a verb not derived from compounding, including NI:
5c) Aydantayattaran → aydann-tayatta-ar-an “He poisoned the well” (lit. “He stored the water with poison).
Type II Noun Incorporation - Case Manipulation
As Mithun identified in Case Manipulation NI, an important function of IN Minhast is to alter the argument structure of a clause. The prototypical function of NI is to decrease the valency of a verb; the transitivity of a clause is decreased by removing one of the core arguments, namely the PT, and absorbing it into the verb. This opens up the Absolutive position to be occupied by another argument, either an oblique argument, or the Ergative argument.
Returning back to Sentence 1a and 1b, the argument structure has been altered from a transitive clause in Sentence 1a to an intransitive one via the application of NI previously observed in Sentence 1b. Both sentences have been restated here as Sentences 5a and 5b:
5a) Yadukte kaslub ayupparu → yaduk=de kaslub ayupp-ar-u (boy=ERG dog.ABS point.at-PST TRANS) “The boy pointed at the dog.” 5b) Yaduk ayukkaslubaran → yaduk ayup-kaslub-ar-an (boy.ABS point.at-dog-PST- INTRANS) “The boy pointed at a/the dog” (lit: “The boy dog-pointed”).
The alteration of the argument structure from a transitive sentence to an intransitive one is pragmatically motivated and changes the nature of the discourse. The incorporation of the PT kaslub, both opens up the Absolutive position for occupation by another argument, in this case, the Agent yaduk. The incorporation of the Patient kaslub also backgrounds it, reducing its salience in the discourse. The Agent yaduk thus becomes more salient, as it has now become the sole core argument of the sentence. The result alters discourse by presenting the Agent as the most important element of the discourse, while that of the Patient has been reduced to a peripheral role.
Patients are not the only arguments that can be subjected to NI. An interesting feature of NI in Minhast is that the semantic nature of a verb may allow certain non-PTs oblique arguments, namely Instrumentals and Locatives, to be incorporated. This is similar to Ainu, another polysynthetic language of Northeast Asia unrelated to Minhast . Like Minhast, Ainu can target Instrumental arguments for NI, like sapa (“head”) underlined below in Sentence 5:
13 This is an instance of Case Manipulation NI, i.e. Mithun's Type II classification. 14 There still remain some die-hard advocates who group the two languages with Chutchki and Nivkh under a single family called Amuran, a reference to the Amur Riverine System. Surface similarities have been ascribed to areal features.
6) Ratki apa a-sapa-e-puni → hung door 1s.A-head-APPL-lift
“I lifted the suspended door with my head.” (J. Runner & Raul Aranovich 2003).
NI of Oblique arguments in transitive clauses does not affect valency, since the Absolutive position remains occupied by the PT. However, the incorporation of Instrumental and Locative arguments modify the meaning of the verb. Speakers are aware that in utilizing this form of NI, they are essentially creating new vocabulary on-the-fly. These new words may be created as one-time entities for the current speech event, or they may be institutionalized and become permanent vocabulary in the lexicon. Examples of Instrumental and Oblique NI are demonstrated in Sentences 6a – 6e:
7a) Yakte dūy kallutaššiakaru → yak=de dūy kallut-haššia-ak-ar-u (1S=ERG salmon.ABS eat- with.chopsticks-3P.INANIM.ABS+1S.ERG-PST-TRANS) – lit: “I eat.with.chopsticks the salmon.”
7b) Yak (dūyaran) kallutaššiekarampi → yak (dūy=aran) kallut-haššia-ek-ar-an-pi (1S.ABS (salmon=DAT) eat.with.chopsticks-1S.ABS-PST-INTRANS-ANTI)
“I eat (some salmon) with chopsticks.”
7c) Yak asunkallutaššiekaran → yak kallut-haššia-ek-ar-an (1S.ABS HAB-eat.with.chopsticks- 1S.ABS-PST-INTRANS)
“I eat/am eating with chopsticks.”
7d) Yak iknatūmanekaran → yak ikna-tūman-ek-ar-an (1S.ABS go-house-1S.ABS-PST- INTRANS)
“I went home.”
7e) Yak ikassuhūrekaran → yak ikassu-hūr-ek-ar-an (1S.ABS rest-mountain-1S.ABS-PST- INTRANS)
“I rested on the mountain.”
The surfacing of the Antipassive in Sentence 7b illustrates that NI of Instrumentals, in this case haššia, does not and cannot decrease valency. Instead, Antipassivation is the only valency-decreasing operation available to reduce the valency of a verb whenever an Instrumental oblique argument is incorporated.
Similarly, the incorporation of a Locative or Goal noun into a positional or locomotive verb does not affect valency, as in Sentence 7d and 7e.
Sentence 7b illustrates that no discussion of Noun Incorporation in Minhast can be made without explaining the interrelation between NI and Antipassivation. Some polysynthetic languages use both NI and Antipassivation to manipulate argument structures. Minhast is among this group of languages. An observation noted among these languages with both NI and Antipassivation is that NI automatically triggers the Antipassive to surface, as demonstrated in Sentences 8a and 8b, taken from Nishga, a North American indigenous tribe of the Pacific Northwest (Mithun, 1984); and Sentences 8c and 8d, taken from Yucatec Mayan (Bricker, 1978):
8a) səməyé:n-sk -m-hó:n → simijeehisgumhoon to.smoke-ANTIPASS-ADJ-fish 'To smoke fish'
8b) íc'l-'sk -m-tá:la → lits'ilsgumdaala to. count. up-ANTIPASS-ADJ-money 'to keep track of
money donated at a potlatch'
8c) t-in-č'ak-∅-ah če'. COMP-I-chop-it-PERF tree “I chopped a tree.”
8d) č'ak-če'-n-ah-en. Chop-tree-ANTIPASS-PERF-I(ABS) “I wood-chopped” = “I chopped wood.”
There are other languages that utilize both NI and Antipassivation where NI does not trigger the Antipassive to surface. The Northeast Asian language Chukchi, a branch of the Chukotko-Kamchatkan is such an example (Kozinsky 1976). Sentence 9a shows NI of an Instrumental noun without Antipassivation. However, Sentence 9b the Antipassive affix appears, but NI is absent because the Antipassive has demoted the PT qora to an oblique Instrumental argument. 9a) Kopalhin na-aldt-koqenat. → walrus.blubber they-knife-mincing “They are mincing walrus blubber with a knife.
9b) Ine-lqerir-ə-rkən (qora-ta). → APASS-seek-PRES-/3sS (deer-INSTR) “He is seeking (for deer).” These two examples demonstrate that NI and Antipassivation are two distinct processes in Chukchi, whereas in Yucatec Mayan they appear to be a unitary process, as they appear simultaneously. Like Chukchi, Minhast treats NI and Antipassivation as two separate, distinct processes. In Minhast, NI and Antipassivation may complement each other, and other times they may be mutually incompatible. Sentences 10a and 10b are presented below. Sentence 10a shows a situation where both NI of an Instrument and the surfacing of the Antipassive affix -pi- occur in the same sentence:
10a) Yak dūyaran kallutaššiekaranampamā, [PRO] yusnakekaran, wassetta. → yak dūy=aran kallut-haššia-ek-ar-an-pi=mā, [PRO] yusnak-ek-ar-an, wa=setta (1S.ABS salmon=DAT eat.with.chopsticks-1S.ABS-PST-INTRANS-ANTI=SUBORD, [PRO] be.late- 1S.ABS-PST-INTRANS, CONN=even.though) “I ate (some) salmon with chopsticks, even though I was late.” As mentioned earlier, incorporation of an Instrumental noun does not decrease valence, so Antipassivation was required to decrease the valency of Sentence 10a. Here, an Instrumental noun, haššia, has already been incorporated into the verb kallut (to eat). Since only one noun can be incorporated into a verb complex at any one time, the only way to reduce the valency of the clause is to apply Antipassivation to demote the semantic Patient dūy (salmon) from core status. The application of 15 The situation with Yucatec Mayan is somewhat more complicated. The language displays split ergativity conditioned by tense: Erg-Abs alignment is observed in the perfect aspect (glossed by Bricker as COMP “completive”), while Nom- Acc alignment is found in the imperfect aspect. This is why the Antipassive does not surface when NI occurs in the imperfect aspect, because Antipassives occur only in the Erg-Abs component of split ergative systems.
the Antipassive marker -pi has demoted dūy from core Absolutive status to a Dative oblique argument marked with the clitic =aran, the typical oblique marker used to mark demoted semantic Patients. Such an operation is required if the speaker wishes to convert yak to Absolutive status for syntactic reasons (e.g. to become the S/O pivot16 to feed into subordinate clauses, as in this case). In contrast, the situation in Sentence 10b shows NI of a Patient. Here, the NI of the Patient blocks Antipassivation: 10b) Yaduk ayukkaslubaran → yaduk ayup-kaslub-ar-an (boy.ABS point.at-dog-PST- INTRANS) – “The boy pointed at a/the dog” (lit: “The boy dog-pointed”). Here, the semantic Patient kaslub (dog) has been incorporated into the verb ayup (to point), but here the Antipassive does not surface. Nor can it without rendering the sentence ungrammatical, because Antipassivation reconfigures a sentence's argument structure by demoting or removing Patients. When kaslub was incorporated, there was no Patient available for the Antipassive to target because the Patient had already been removed from core status by NI. To illustrate the reason why Antipassivation was blocked by NI of a Patient, Sentence 10c presents a theoretical situation resulting from the application of Antipassivation after a Patient has been incorporated: 10c) **Yadukaran [missing ABS] ayukkaslubarampi → yaduk=aran [missing ABS] ayup- kaslub-ar-an-pi (boy.ABS point.at-dog-PST-INTRANS-ANTI)
Sentence 10c is ungrammatical because NI has already removed the PT (kaslub), resulting in a monovalent clause with only one core argument, the Absolutive (yaduk). Antipassivation would
Sap min Aynuw-ast min gubbāturrād sattabe-
demote the Absolutive yaduk to a Dative peripheral argument, yadukaran, leaving no other argument to 17 fill in the missing Absolutive position. Minhast forbids zero-valent sentences , so the simultaneous application of NI and Antipassivation renders Sentence 10c infelicitous. Thus, both Minhast and Chukchi treat NI and Antipassivation as two separate, distinct operations. But whereas the Chukchi samples showed that the Antipassive does not surface when even an Instrumental noun is incorporated, Minhast allows Antipassivation of Instruments if valency reduction to demote a Patient is required, allowable provided the IN is an Instrumental argument. 16 A pivot is a syntactic construct in multi-clause sentences wherein the syntactic role of a core argument that has been omitted by PRO-drop, is recovered from its role in a higher level clause. There are two types of Pivots identified by Robert Dixon in Ergativity (1994): S/A Pivots and S/O Pivots. S/A Pivots, associated with Nominative-Accusative language, assign the Nominative argument as the argument (Pivot) that is coreferrent with the null argument of successive clauses. Syntactically Ergative languages, in contrast, assign the Absolutive as the Pivot of successive clauses. 17 There is one exception to this rule: Minhast has a closed set of Interrogative verbs, some of which take zero core arguments. -ar-an- =naft yak=de nusill-ek-u, ...[PRO
= gubbāturrād] nurr-wastane-0-ek-nes-u-d
“Sap min Aynuw-ast min gubbāturrād
sattabe-redad-ar-an=naft yak=de nusill-ek-u”
The use of Case Manipulation accounts for the majority of NI in Minhast. This is not surprising, since Minhast, as a syntactically ergative language, utilizes various grammatical devices to maintain and manipulate the S/O pivot to cross-reference the Absolutive argument across clauses.
Type III Noun Incorporation - Discourse Manipulation
Another important difference in Minhast between Antipassivation and Noun Incorporation is found in discourse manipulation in extended speech or narratives. Antipassivation is often used to remove an element from discourse entirely, marking the demoted NP as truly incidental and ultimately unimportant to the narrative. On the other hand, Minhast uses NI to retain the demoted argument in the 18 discourse, albeit backgrounded . Backgrounding is a useful device if the salience of the demoted argument must be decreased, but is still too important to eliminate from the discourse. Therefore, an IN may be retained throughout the length of the narrative. In doing so, the speaker can maintain the context in which the core arguments operate, a useful mnemonic and rhetorical tool to keep the listener 19 on track with the discourse . Additionally, the IN may be promoted back to a core argument later on without creating an abrupt discontinuity within the narrative, which would occur if an argument that had been supposedly marked incidental in the discourse by Antipassivation were to be reintroduced into the argument. Indeed, it is the case that speakers often expect a noun backgrounded by NI in extended discourse to later resume core status at a later point. The following sample passage comes from a recording of a Salmon Speaker conversation, and illustrates the backgrounding function of NI: 11) Ayakna ubarrāsekabampamā, dūyaran ehahekampi, kašmakki ehaheku, arruššan sakwaštan, yakmaran saxtihayyuštinesaš, saxtimurrešnesašnaft irraktuyeku, wahēk intušindūyekan, ayayadūyekan, indiknatūmanekammā, hūraran indiħyekan, reyriaktipār sipsambuyekammā, iyuššitaran sarekampamā, intadutyusapakmahummā, dumyaharan sarekampi, yapkekan, uzzattuyekan, bastettuyekanimmā, isangarekan. Wahēk, uzzatekunimmā, inditayyisakšareku, alluakan. Gloss: Ayakna ubarrās-ek-ab-an-pi=mā dūy=aran ehah-ek-an-pi, kašmak=ki ehah-ek-u, arrusi-š-an sakwat-š-an, yakma=ran saxt-hayyut-š-nes-an=š, saxt-murre-š-nes-an=š=naft irrak-dūy-ek-u, wahēk nta-ušn-dūy-ek-an, ayaya-dūy-ek-an, nd-ikna-tūman-ek-an=mā, hūr=aran nd-hy-ek-an reyria-k-de=pār sipsab-dūy-ek-an=mā, iyuššit=aran sar-ek-an-pi, nta- dut-yusap-akmah-u=mā, dumyaħ=aran sar-ek-an-pi, yapk-ek-an, uzzat-dūy-ek-an, bastet-dūy- ek-an-nimmā, isangar-ek-an. Wahēk, uzzat-ek-u-nimmā, n-tayy-sakšar-ek-u, alluak-an. early.morning to.fish-1S.ABS-IMPF-INTRANS-ANTI=SUBORD salmon=DAT capture- 1S.ABS-INTRANS-ANTI, kašmak.river=LOC capture-1S.ERG-TRANS, be.big-ANIM- INTRANS be.fat-ANIM-INTRANS, salmon.noodle.soup=DAT INCHO-dinner-ANIM.ABS- IMMED.FUT-INTRANS=IRREAL INCHO-be.delicious-ANIM-IMMED.FUT- INTRANS=IRREAL=NMLZ know.general.truth-salmon-3S.COMM.ABS + 1S.ERG-TRANS, wahēk INTENS-hit-salmon-1S.ABS.INTRANS, INCEP-go-home-1S.ABS- INTRANS=SUBORD mountain=DAT INCEP-be.high-1S.ABS-INTRANS shoulder-1S.ERG- ERG=INSTR carry-salmon-1S.ABS-INTRANS=SUBORD, storm.clouds= DAT see-1S.ABS- INTRANS-ANTI INTENS-DAT.APPL-rain-1S.ABS+3P.ERG-TRANS=SUBORD, cave=DAT see-1S.ABS-INTRANS-ANTI, take.shelter.from.bad.weather-1S.ABS-INTRANS, pull.out.of.container-salmon-1S.ABS-INTRANS, place.down.carefully-salmon-1S.ABS- INTRANS=PURPOS.SUBORD light.up.a.fire-1S.ABS-INTRANS . wahēk, 18 Discourse Manipulation falls under Mithun's Type III classification. But the discourse manipulation function of backgrounding information in Minhast is very different from other languages with NI, especially the oft-cited works on Mohawk NI processes. These differences are described in greater detail later in the text. 19 The use of NI to background previously mentioned items falls under Mithun's Type III Noun Incorporation category.
pull.out.of.container-3S.COMMON.ABS+1S.ERG-TRANS-PURPOS.SUBORD INCEP- prepare.food-REFL.BEN-3S.ABS+1S.ERG-TRANS, be.pregnant.with.eggs-INTRANS. “Early in the morning, I was fishing and caught a salmon in the Kashmak River. It was a big, fat one, and I knew it would make good yakma for dinner. So I killed it and put it in my basket, and I began my journey home, heading up the mountain carrying the salmon over my shoulder, but then I saw storm-clouds and soon the rains began to pelt me. I found a small cave and took shelter in it, and then I took out the salmon and lay it aside so I could make a fire. I took the salmon out and began to clean it, it was a female and filled with roe.” The passage introduces the salmon in the beginning: “...I was fishing and caught a salmon...” Here, the salmon, while semantically Patient, is introduced in the narrative as a Dative argument, and becomes the topic of interest in the next few clauses, where it assumes core status as the Absolutive argument. Casting the salmon as an Absolutive argument also ensures that the sentence fulfills the requirements for its Absolutive argument to become the S/O pivot of the next clause, which is a stative clause. Minhast lacks a separate word class for adjectives, hence stative verbs are used instead. Statives target only Absolutives, never Ergatives or Obliques. After the salmon is killed and placed in the basket, the wahēk particle20 allows the narrator to resume his status as the Absolutive argument so that he can serve as the S/O pivot for the next series of clauses up to the point in the narrative where he makes a fire. Throughout this part of the narrative, the salmon, when mentioned, has been backgrounded by NI so that the narrator can retain his status as the S/O pivot. The second wahēk of the narrative in the next clause allows the narrator to change the S/O pivot again, at which point he reassigns the salmon to Absolutive status in order for it to become the S/O pivot to feed the final three clauses of the narrative. The same native speaker who provided this narrative was presented an alternative text, whose English translation is exactly the same, but instead of NI to demote the salmon and move it into the background, Antipassivation was used instead to demote the salmon to a Dative oblique argument: 12) Ayakna ubarrāsekabampamā dūyaran ehahekampi, kašmakki ehaheku, arruššan sakwaštan, yakmaran saxtihayyuštinesaš saxtimurrešnesašnaft irakeku, wahēk dūyaran intušnekampi, dūyaran ayayakampi, indiknatūmanekammā hūraran indiħyekan reyriaktipār dūyaran sipsabekampamā, iyuššitaran sarekampamā intadutyusapakmahummā, dumyaharan sarekampi, yapkekan, dūyaran uzzatekampi, dūyaran bastetekampinimmā isangarekan. Uzzatekunimmā inditayyisakšareku, alluakan. Gloss: Ayakna ubarrās-ek-ab-an-pi=mā dūy=aran ehah-ek-an-pi, kašmak=ki ehah-ek-u, arrusi-š-an sakwat-š-an, yakma=aran saxt-hayyu-š-nes-an=š saxt-murre-š-nes-an=š=naft irak- ek-u, wahēk dūy=aran nt=ušn=ek=an-pi, dūy=aran ayaya-ek-an-pi, nd-ikna-tūman-ek-an=mā hūr=aran nd-ħy-ek-an reyria-ek=de=pār dūy=aran sipsab-ek-an-pi=mā, iyuššit=aran sar-ek- an-pi=mā nta-dut-yusap-s-ek-u=mā, dumyah=aran sar-ek-an-pi, yapk-ek-an, dūy=aran uzzat- ek-an-pi, dūy=aran bastet-ek-an-pi=nimmā isangar-ek-an. Uzzat-ek-u=nimmā nd-tayy-sakšar- ek-u, alluak-an. early.morning to.fish-1S.ABS-IMPF-INTRANS-ANTI=SUBORD salmon=DAT capture- 1S.ABS-INTRANS-ANTI, kašmak.river=LOC capture-1S.ERG-TRANS, be.big-ANIM- 20 The wahēk particle performs multiple functions. It is a type of subordinative sequential particle roughly translated as “and then”. It differs from the -mā suffix in that it simultaneously allows the Absolutive and Ergative arguments to switch roles. The switch in roles between the core arguments coincides with a change in topic, so wahēk serves as a topic shifter as well.
INTRANS be.fat-ANIM-INTRANS, salmon.noodle.soup=DAT INCHO-dinner-ANIM.ABS- IMMED.FUT-INTRANS=IRREAL INCHO-be.delicious-ANIM-IMMED.FUT- INTRANS=IRREAL=NMLZ salmon=DAT know.general.truth-3S.COMM.ABS + 1S.ERG- TRANS, wahēk salmon=DAT INTENS-hit-1S.ABS.INTRANS-ANTI, INCEP-go-home- 1S.ABS-INTRANS=SUBORD mountain=DAT INCEP-be.high-1S.ABS-INTRANS shoulder- 1S.ERG-ERG=INSTR salmon=DAT carry-1S.ABS-INTRANS-ANTI=SUBORD, storm.clouds= DAT see-1S.ABS-INTRANS-ANTI INTENS-DAT.APPL-rain- 1S.ABS+3P.ERG-TRANS=SUBORD, cave=DAT see-1S.ABS-INTRANS-ANTI, take.shelter.from.bad.weather-1S.ABS-INTRANS, salmon=DAT pull.out.of.container-1S.ABS- INTRANS-ANTI, salmon=DAT place.down.carefully-1S.ABS-INTRANS- ANTI=PURPOS.SUBORD light.up.a.fire-1S.ABS-INTRANS. wahēk, pull.out.of.container- 3S.COMMON.ABS+1S.ERG-TRANS-PURPOS.SUBORD INCEP-prepare.food-REFL.BEN- 3S.ABS+1S.ERG-TRANS, be.pregnant.with.eggs-INTRANS. When asked about the alternate narrative, the consultant said that while it was grammatically well- formed, he had not expected to hear it mentioned multiple times throughout the narrative. He remarked that the salmon had been demoted by the first occurrence of the Antipassive, and asked why the Antipassive was used if the salmon was important enough to remain a part of the narrative until it would later resume its core status. In this respect, NI in Minhast differs in its discourse function from that of Mohawk. An IN in extended Mohawk discourse, as backgrounded information, is truly incidental information, meant only to provide contextual information in the body of the discourse, as illustrated in this passage (Mithun:1984): 13) ó:nv yeyóhe n-a-ye-nvhst-ayvtho:-ko. Akwe: tsi t-ka-nvhst-ayvth-u yvyakwe' tanu y-v-yak- wa-hroht-v-ht-e. E-tho ne ó:nv v-yak-wa-nor-oht-hsi-'. → then there.it.set there-would-one- corn-plant-REVERS all to there-it-corn-planted will.we.go and there-will-we-all-ear-fall- CAUSE-PUNC then the then will-we-all-husk-stand-REVERS-PUNC “Then it was time to harvest it (the corn). We would all go to the cornfield and take it (the corn) from the stalks. We would then husk it.” While Minhast uses NI for backgrounding purposes as in Mohawk, its motivation for backgrounding an entity through NI is significantly different from Mohawk and the rest of the Iroquioan languages. Minhast employs NI to keep the IN available so that it can later restore it to core status, thereby returning it to the foreground in the discourse. Thus, in Minhast the status of an IN is somewhere between an oblique argument and a core argument, a key difference between how it uses NI as opposed to Mohawk. The application of NI in Minhast serves as a foreshadowing technique; temporarily backgrounded, the IN often resurfaces later in the discourse as a principal player (core argument) of the narrative. The foreshadowing function of NI also demonstrates an important point: INs occurring in 21
extended discourse are highly referential in Mohawk, as argued by Baker (1996). Minhast, based on native speakers’ self-reports.
The same is true in Additionally, there is a type of Control clause, namely the Causative clause22 which requires the 21 This point seems to be contested by others, at least in the case of Mohawk. See Mithun 2010:12. 22 In addition to Causative clauses, Control clauses also include Cognitive clauses, Desiderative clauses, Mandative clauses, Decision clauses etc. Causative clauses differ from other types of Control clauses in that Causative clauses are also Previous Event-Successive Event conjoined clauses (PEC-SEC), always recognizable by the Subordinator affix -
combination of NI and Antipassivation if the speaker wishes to reduce valency. As with all Control clauses, Causative clauses are highly transitive and increase valency. As in all Control clauses, they are recognizable by the presence of two subjects, one called the Controller, which manipulates or acts on a secondary subject, called the Controllee. The Controllee is responsible for carrying out the actual event or achieving the final end state. Controllers are typically Ergative arguments, and Controllees are cast to Absolutive status, as Sentence 13a demonstrates: 13a) Yakte anxekte išpisattabeddadekarumā siyyekittūmarraran . Yakte šp-sattabe-redad-ek-ar-u=mā siyyekit-tūman-ar-an 1S.ERG brother.ABS CAUS-kill-man-1S.ERG+3S.ABS-PST-TRANS=SUBORD burn-house- PST-INTRANS) “I sent my brother to kill the man and then he (i.e. the brother) burned down the house.” Sentence 13a targets the brother as the Controllee, and so was cast as an Absolutive. However, sometimes valence-reducing operations are required to convert or retain the Controller as the S/O pivot. This can be done only by demoting or removing the Controllee23 acting on the semantic Patient. In such a case, the logical PT is incorporated into the verb in the matrix. The Controllee can now be removed from the Absolutive position, leaving it vacant for the Controller to move into the vacated Absolutive space. The combination of NI and Antipassivation yields the example in Sentence 11b: 13b) Yak išpisattabeddadekarampamā siyyekkittūmanekarampi. Yak šp-sattabe-redad-ek-ar-an-pi=mā [PRO] siyyekkit-tūman-ek-ar-an. S.ABS CAUS-kill-man-1S.ABS-PST-INTRANS=SUBORD burn-house-1S.ABS-INTRANS “I sent someone to kill the man, and then I burned down his house.” The Absolutive argument, yak, is now available to feed the S/O pivot, thanks to the coordination of both NI and Antipassivation to reduce the valency of the sentence; NI backgrounds the logical PT redad, and the application of Antipassivation allows yak to occupy the Absolutive position. Yak can thus serve as the Absolutive feeder of an S/O pivot. Other Control clauses differ from Causative clauses in that they are embedded in their matrix clauses as 24 complements, i.e. nominalizations marked with the clitic =naft . Cognitive and Perception clauses are mā joining the PEC to the SEC. 23 The Causative Controllee can be brought back into the clause, but it must appear as an Instrumental oblique argument. The reading could then be calqued: “By using [NP], I caused to kill the man so that I may live.” Languages that have the Passive voice, such as English, often use an Instrumental oblique marker on a demoted subject just as languages with the Antipassive do with Instrumental oblique arguments. Compare English “I was caused by Joe to kill the man so that I could live” with Minhast “Yak anxē=par išpisattabeddadekarampamā siyyekkittūmanekaran”, lit. “By using my brother, I caused to kill the man and then I burned down the house”, where the Controllee, anxea (“brother”), is retained as a demoted subject via Antipassivation. 24 The use of nominalizations to serve as complements can lead to ambiguities: Sap min Aynuwast gubbāturrād sattabeddadaranaft yakte sarekaru can be interpreted as either “I saw this Ainu soldier who killed the man” or “I saw
included in this subtype: 14a) Sap min Aynuwast min gubbāturrād sattaberedadaranaft yakde nusilleku 25 Sap min Aynuw-ast min gubbāturrād sattabe-redad-ar-an=naft yak=de nusill-ek-u. DEM.PROX CONN aynu.person-GENT CONN warrior kill-man-PST-INTRANS=NMLZ 1S=ERG know.a.fact-3S.ABS + 1S.ERG-TRANS “I know that this Ainu warrior killed the man” or “I know this Ainu warrior who killed the man.” 14a.1) Sattabe-redad-ar-an=naft, yak=de sap min Aynuw-ast min gubbāturrād nusill-ek-u. “ 14b) u. 14c) Yak nusill-ek-an-pi. “I know a fact [about something]”. 14d) Yak nusill-redad-ek-an-pi. “I know a fact [regarding something about] the man.” 14e) Sap min Aynuwast min gubbāturrād sattabe-redad-ar-an=naft yak=de sar-kaddara-ek-ar- u. “I event-saw him, this Ainu warrior killed the man.” 14e.1) Sattabe-redad-ar-an=naft, yak=de sap min Aynuwast min gubbāturrād sar-kaddara-ek- ar-u. “I event-saw this Ainu warrior, he that killed the man.” 14f) Redad=aran sap min Aynuwast min gubbāturrād sattabe-ar-an-pi=naft yak=de sar- kaddara-ek-ar-u. “I event-saw him, this Ainu warrior killed a man” or “I event-saw him, this Ainu warrior who killed a man”. 14f.1) Redad=aran sattabe-ar-an-pi=naft yak=de sap min Aynuwast min gubbāturrād sar- kaddara-ek-ar-u. “I event-saw this Ainu soldier, he that killed a man”. 14g) Yak sar-ek-ar-an-pi. “I saw [something].” 14h) that this Ainu soldier killed the man.” With non-third person pronouns this ambiguity disappears: Sattabeddadtaharanaft yakte sartahekaru (lit: “I saw you that killed the man”) can only be interpreted as a complement. Another strategy for disambiguation is NI of -kaddar(a)- “event” or -razn- “matter, issue, subject” to the Control verb, e.g Sap min Aynuwast gubbāturrād sattabeddadaranaft yakte sar-kaddara-ekaru (“lit: “I event-saw the Ainu who/that killed the man.”). However, NI of -kaddar(a)- and -razn- are optional, and more often than not omitted. .
kill-man-PST-INTRANS=NMLZ 1S=ERG DEM.PROX CONN aynu.person-GENT know.a.fact-3S.ABS + 1S.ERG-TRANS He who killed the man, I know this Aynu warrior (did it)” or “I know this Ainu warrior who killed the man.” Redad=aran sap min Aynuw-ast min gubbāturrād sattabe-ar-an-pi=naft yak=de nusill-ek- “I know that this Ainu warrior killed a man” or “I know this Ainu warrior who killed a man.” 25 The verb nusill- alleviates the ambiguities mentioned in footnote #17, since its full meaning is “to know a fact”.
a) ** Yak sar-redad-ek-ar-an-pi. “I man-saw.” Antipassivation + NI is not allowed here. b) Anxea yak sar-redad-ek-ar-an-pi. “I man-saw (my) brother.” Antipassivation + NI is allowed here because this is Type IV NI (see below) 14i) Sap min Aynuwast min gubbāturrād sattabe-redad-ar-an=š=naft yak=de ittawas-ek-ar-u. 14j) Redad=aran sap min Aynuwast min gubbāturrād sattabe-ar-an-pi=naft yak=de ittawas-ek- u. 14k) Yak ittawas-ek-ar-an-pi. 14l) **Yak ittawas-redad-ek-ar-an-pi. Notice that if the Controller is placed into the Absolutive position by Antipassivation (Sentences 14d, 14h, 14l) the nominalization must be deleted, since Minhast does not tolerate two separate Absolutive arguments and nominalizations must always have an Absolutive argument to relativize on. The IN of the complement clause in which the Controllee is embedded in may be retained by incorporation into the Control verb, but without nominalization, there is no way to qualify the NI. That must be determined by context.
Type IV Noun Incorporation - Classificatory Functions
Some languages have expanded the range of functions that NI can perform beyond valence operations and discourse pragmatics. These languages take NI to an advanced level, whereby a generalized noun is incorporated to classify or categorize the Patient, which has more specific meaning. Again, Mohawk is an exemplary language that exploits this form of NI pervasively, called Classificatory NI (Mithun's Type IV category). The following passage from Mithun (1984) illustrates this NI function type: 14a) Tohka niyohserd:ke tsi nahe' sha'te:ku niku':ti rabahbot wahu-tsy-ahni:nu ki rake'niha. → several so.it.year.numbers so it.goes eight of.them bullhead he-fish-bought this my.father. “Several years ago, my father bought eight bullheads.” In this example, the incorporated noun (i)tsy (“fish”) is co-referrent with the Patient rabahbot (“bullhead fish”); (i)tsy serves as a general modifier to classify the more specific rabahbot. As the narration continues, rabahbot is subsequently referred to by the incorporated noun (i)tsy: 14b) Saháhkete' ki:kv rakenuhá:'a s-a-h-vtsy-ahsheruny-à:na-'. Yusa :rawe ki': óksa'k wa-h-vtsy- ahserul:ni tanu wa-h-vtsy-akeri:tahw-e. Tsi n-a-ho-tsy-ari-hs-e ki' ki:kv wahv:ru, “Tho yukyatv:ro rinu-tsy-anut-v-:ra.” back.he.turned this my.uncle back-PAST-he-fish-fix-go.to-PUNC back-he-arrived just quick PAST-he-fish-fix(PUNC) and PAST-he-fish-fry-PUNC as as-PAST-him-fish-fry-finish-PUNC just this he.said there we.two.friends.are I/him-fish-feed-for-go.to “My uncle then returned to fix them (the fish). At home, he cleaned and fried them (the fish), and when they (the fish) were ready, he decided to take them (the fish) over to his friend as a special treat.” The Mohawk examples show how the generic IN, (i)tsy, is used like an agreement marker throughout the entire discourse. Although it is generic, it is used to refer back repeatedly to the more specific rabahbot (bullheads), which is mentioned only once, at the beginning of the narrative. Minhast, like Mohawk, uses Classificatory NI, but once again, the factors motivating Minhast NI to implement Classificatory NI are different. As mentioned earlier, certain noun classes are resistant to NI or forbidden. Nouns that are resistant to NI or forbid it are typically located in the high end of the animacy scale. Nouns that fall under this portion of the animacy hierarchy are kinship nouns and proper nouns. As in Mohawk, for Classificatory NI Minhast incorporates a generic noun. This incorporated argument is then used as an agreement marker to the NI-resistant noun, which is always in the Absolutive. A requirement for Classificatory NI in Minhast is that the noun which the IN is corefererent with cannot be a semantic Agent, and for this reason, once again Antipassivation is blocked by NI. The NI-resistant noun and the generic IN that serves as its agreement marker are highlighted in bold underlined font: 15a) Supnar min anxekte Keyyis min niħkašektaran duntittarraru duntittarumā, anxēa indintanuskikirimredadnnarumā. Supnar min anxē=ek=de Keyyis min niħkaš-ek=de=aran dut-nittam-ar-u=mā, anxēa nd-inta- nusk-kirim-redad-nn-ar-u=mā... (PROP.NOUN CONN brother=3MS.ABS+1S.ERG=ERG PROP.NOUN CONN friend-3S.ABS+1S.ERG=ERG=DAT DAT.APPL-shout-PST- TRANS=SUBORD, brother.ABS INCEP-INTENS-ADVER.APPL-speak-man- 1P.EXCL.ERG-PST-TRANS=SUBORD) “My brother Supnar shouted at my friend Keyyis and we started to argue against my brother...” The narrative continues. The narrator decides to cast himself as the Absolutive argument to feed the S/ O pivots in the succeeding clauses. When he needs to mention his brother as the Patient, he chooses not to use the Antipassive in order to cast his brother as a Dative oblique; to do so would imply the brother is incidental information, which is not the case in this passage. Instead, he applies NI on redad, which coindexes his brother as the logical Patient. Redad then serves as a proxy for his brother for NI operations: 15b) ...kūdāš segwekarammā, (kūde) yummatekarumā, rabbaddadekarammā, karyaħtendepār kaħmadekarampamā šarrataran ušniddadekarammā, nittarredadekarannamā: “Bakran wattaħte ušnktahuš? Hatā anxēšattarakš? Ta'astakkemarunaft wastānešattarakte hittastānehakkemaruš?” ...kū=dāš segw-ek-ar-an=mā, (kua=de) yummat-ek-ar-u=mā, rabba-redad-ek-ar-an=mā, karyaħt-enn=de=pār kaħmad-ek-ar-an-pi=mā šarrat=aran ušn-redad-ek-ar-an=mā, nittam- redad-ek-ar-an=namā: “Bakran wa=tah=de ušn-ktah-u=š? Hatā anxēa-šattar-hak=š? Ta=ast-hakkem-ar-u=naft wastāne-šattar-hak=de hitt-wastāne-hakkem-ar-u=š?” 3S.OBL=MAL make.fist-1S.ABS-PST-INTRANS=SUBORD, (3S=ERG) swing.fist- 1S.ABS+3S.ERG-PST-TRANS=SUBORD, grab-man-1S.ABS-PST-INTRANS=SUBORD, arm-3S.NEUT.ANIM.ABS+3S.ERG=ERG=INSTR twist-1S.ABS-PST-INTRANS- ANTI=SUBORD earth=DAT hit-man-1S.ABS-PST-INTRANS =SUBORD, shout.at-man- 1S.ABS-PST-INTRANS=DIR.QUOT: why CONN=2S=ERG hit-1S.ABS+2S.ERG- TRANS=IRREAL? NEG brother-RECIP-1P.INCL.ABS=IRREAL? NEG= beget- 1P.INCL.ABS+3P.COMMON.ERG-PST-TRANS=NMLZ blood-RECIP - 3S.NEUT.ANIM.ABS+1P.INCL.ERG=ERG give-blood-1P.INCL.ABS+3P.COMMON.ERG-
PST-TRANS=IRREAL “...[and] we started to fight, my brother and I. He swung at me (but missed), and I grabbed my brother, then I twisted his arm and struck him to the ground. I yelled at my brother, 'Why do you strike at me? Am I not your brother? Are we not of the same blood? (lit. Our mutually- shared blood, did not those that begat us give us blood?)'” While these passages demonstrate that Minhast does employ Classificatory NI, it does not do so as extensively as in Mohawk. In Minhast, Classificatory NI is employed to get around the obstacles presented by NI-resistant nouns, which by their nature, rank high in the animacy hierarchy. Thus, Classificatory NI is rarely, if ever, encountered in passages with low-animacy entities. Most important, though, is whether Classificatory NI is required to maintain the S/O pivot. If not, Classificatory NI is not employed. Thus, one is not likely to see a sentence like in 16a. In fact, a native speaker would find it quite odd, and most likely ungrammatical: 16a) Dūy zaydakkī sarekaru, (dūy) rimar-sankūy-ekarumā, (dūy) niyyet-sankuy-ekaru. “I saw the salmon in the river, I fish-speared it, then I fish-pulled it out of the water.” The sentence does not require Classificatory NI because dūy (“salmon”) is incorporatable and no advantage is gained by using the generic noun sankūy (“fish”) to co-index dūy. While both nouns are animate, they are also neuter in gender, which when compared to nouns that have masculine and feminine nouns, both of these nouns lie lower in the animacy spectrum; Classificatory NI is typically employed with highly animate nouns, which include proper nouns and kinship terms. However, dūy is low in the animacy scale, so the incorporation of the more generic term sankūy for Classificatory NI is unnecessary. Compare this with the previous Mohawk example of Classificatory NI: 16a) Tohka niyohserd:ke tsi nahe' sha'te:ku niku':ti rabahbot wahu-tsy-ahni:nu ki rake'niha. several so.it.year.numbers so it.goes eight of.them bullhead he-fish-bought this my.father. Here rabahbot (the bullhead fish) and generic, incorporated noun tsy (fish) lie in the same level on the animacy hierarchy, but Mohawk utilizes Classificatory NI so that tsy can function as an agreement marker for rabahbot later discourse. In the case of Minhast, the application of Classificatory NI background dūy conveys no observable benefits, as both dūy and sankūy are at the same animacy level. Minhast uses Classificatory NI when the target noun is unincorporatable, as in the case of proper nouns and kinship nouns. However, neither dūy nor sankūy are considered ineligible for incorporation. If dūy is the topic of interest, it must be cast as the Pivot, which requires that it assume the role as an Absolutive argument: 16b) Dūy zaydakkī sarekaru, [PRO] rimarekarumā, [PRO] niyyetekaru. “I saw the salmon in the river, speared it, and pulled it out of the water.”
Or if the pronoun yak/-ek- were the topic of interest, it would be the S/O pivot, and then dūy can become the incorporated Patient: 16c) Dūy=aran zaydakkī sarekarampi, rimar-dūy-ekammā, [PRO] niyyet-dūy-ekan. “I saw some salmon in the river, I salmon-speared, then salmon-pulled-out-of-the-water. ...or dūy can be nominalized, in which it would again serve as the S/O pivot: 16d) Dūy zaydakkī sarekarunaft rimarekarumā [PRO] niyyetekaru. “The salmon that I saw, I speared and pulled it out of the water.” However, the incorporation of dūy would make it ineligible to function as the Pivot. The following
- Sartuyekarammā rimarekaru >> ** sar-duy-ek-ar-an-mā rimar-ek-ar-u
- “I fish-saw then speared it.”
Here the intent of the speaker was to mark duy as a PT in the first clause via NI, then link the first clause to the second clause via the -mā suffix. The verb in the first clause is intransitive after the incorporation process, which is why it received intransitive marking with the -an- affix. The second clause is marked as a transitive verb by the suffix -u. The speaker used the transitive -u suffix thinking that its PT argument was the incorporated -duy-, however this is ungrammatical. The transitive marker requires an Absolutive argument, but the incorporation of -duy- removed it from the Absolutive position, thus there no longer is an Absolutive argument available to rimarekaru to function as the Pivot. In Minhast, an IN can never serve as a Pivot in clause chains, only an explicit NP in the Absolutive case frame, or null-marked implicit NP, whose agreement marker is the ABS agreement marker inside the verb complex. To make this sentence grammatical while retaining the IN, at the minimum an Applicative affix marking an explicit or implicit NP that has been promoted to core status 26 The fact that animacy conditions have to be met before Classificatory NI is applied explains a major reason why it is not as extensively used in Minhast as it is in Mohawk. The requirements of the S/O pivot, which is exploited pervasively in Minhast, also has a role in governing whether Classificatory NI is employed. Classificatory NI appears to be a last-resort measure that is used if no other valence operation can feed the S/O pivot with the proper argument; otherwise, it is not employed. That Minhast has other syntactic mechanisms, such as the valence operations of Antipassivation and Applicative Formation, as well as nominalizations and the wa-clause connective and allied structures all contribute to the low frequency of Classificatory NI in Minhast. The situation for Mohawk is the inverse of the Minhast system: lacking the ancillary syntactic mechanisms for altering argument structure found in Minhast, Mohawk needs Classificatory NI because the language does not employ nominalizations. Mohawk clauses, even when chained together into long clause chains, are still ultimately regarded as truly independent, stand-alone sentences; stripping one of the component clauses from its matrix clause does not change its grammaticality. In 26 The 3S.ABS+1S.ERG agreement marker is -ek-, which is homophonous with 1S.ABS -ek-. The Transitive verb suffix - u disambiguates which -ek- form is being used. is required.
contrast, nominalizations cannot stand alone, they must be contained within a matrix clause; stripping them from their matrix clause does make them ungrammatical. Minhast has its S/O pivot system available to it, which it exploits aggressively, especially in the case of nominalizations; Mohawk has no such syntactic construct. That Mohawk has the Classificatory NI available to it to handle the narratives in 14a and 14b does not make it superior to Minhast. That Minhast can use nominalizations and its S/O pivot system to handle the same narrative and minimize using Classificatory NI does not make it superior to Mohawk. They are simply different syntactic structures available to their respective languages, and each language maximizes the tools available to solve syntactic problems such as the ones just described.
- This is an example of Mithun's Classificatory NI (Class IV) being exploited by Minhast. Here, it is essentially creating the equivalent of a locative noun in other languages, e.g. English "Within the interior of the beast..." Here, nua means "side" has been incorporated into the verb complex. The implicit head is suharak (deerskin), which was mentioned in a previous line in the passage, referred to by the Locative applicative naħk-. This construction is equivalent to saying "Next to it", "By its side", etc.
- Tayyamakim tayyapte suharak naħkixripuxnutartimmahabu
tayyamak min tayyap=de suharak naħk-xr-pux-nua-tar-timmah-ab-u
thunder CONN ball=ERG deerskin LOC.APPL-ITER-boom-side-3S.ANIM.ABS+3S.ANIM.ERG-IMPF-TRANS
Canonballs explode next to it the deerskin.
Truncation/Weak Suppletion
Most Minhast nouns are irregular in their IN forms, the majority of which exhibit what has been referred to among Minhast linguists as truncation, wherein the incorporated nominal appears in a shortened form from the reduction or loss of syllables. This process is more commonly referred to as weak suppletion (see also "Noun Incorporation: A New Theoretical Perspective", Alessio Muro, 2009). Cross-linguistically this process is seen in other languages such as Sora, an unrelated language from the Munda family in India. The first example shows the analytic version of the Sora sentence "Will they eat the buffalo/ Do they eat buffalo?". The second example shows the noun incorporated-version of the same sentence, where the independent word bɔŋtɛl ("buffalo") has lost its final syllable to create its incorporating form, bɔŋ:
- bɔŋtɛlәnәdɔŋ jomtɛji pɔ
bɔŋtɛl-әn-әdɔŋ jom-t-ɛ-ji pɔ
buffalo-/әn/3-ACC eat-NPST-3S-PL.S Q
Will they eat the buffalo/ Do they eat buffalo?
- jombɔŋtɛnji pɔ
jom-bɔŋ-t-ɛ-n-ji pɔ
eat-buffalo-NPST-3S-INTR-PL.S Q
Will they eat the buffalo/ Do they eat buffalo?
Similarly, Minhast INs exhibit weak suppletion, and it occurs extensively, particularly with nouns longer than two syllables, e.g. sussagarānī > -suggan- ("big toe"). The contrast can be seen in the following two examples, the first where the noun occurs in its full form as the dependent argument in a possessive NP, and the second wherein the noun appears in truncated form after noun incorporation:
- Sussagarānītirektiki kahušnišattekaran.
sussagarānī-tirek=de=ki kah-ušn-šatt-ek-ar-an
big.toe-3SN.INAN.POSSM+1S.POSSR=ERG=LOC INV.VOL-hit-RFLX-1S.NOM-PST-TRNS
I stubbed my big toe.
- Kahušnisuggašnattekaran.
kah-ušn-sussagarānī-šatt-ek-ar-an
INV.VOL-hit-toe-RFLX-1S.NOM-PST-TRNS
I stubbed my big toe.
Similar patterns can be found with hispawak > -hispak- ("birch"), and izzesparak > -spark- ("canoe").
The pattern of truncation is unpredictable; syllable loss may occur in initial, medial, or final positions, although noun roots with more than two syllables tend to lose either their medial or final syllables and retain the initial syllable, but exceptions abound, such as allāga > -lgagg- (conch) .
Noun Incorporation of Oblique Arguments
In many languages which exhibit noun incorporation, the type of noun that can be incorporated into the verb is often restricted. Some languages incorporate body parts only, others are restricted to inalienable nouns or some other semantic category. Other languages that exhibit extensive noun incorporation, of which Mohawk and its relatives in the Iroquoian language family are the most studied, while having much fewer semantic restrictions, still limit the syntactic or thematic role of the noun that can be incorporated: these are that of the Patient argument, and in some cases the Instrument argument. Other arguments serving in a different thematic/theta role are barred from incorporation. Other noun incorporating languages, such as Chukchi, appear to have no restrictions on the theta role of the incorporated noun (IN); but when these oblique nouns are incorporated, the only way to recover their thematic role is by context alone.
Minhast is one of those languages that can incorporate oblique arguments. However, the oblique arguments that can be incorporated are constrained by the semantic characteristics of the verb. Some transitive verbs which require a third argument, which is always an oblique noun, can optionally incorporate the oblique noun. The verb wasaskiyu is such an example:
wasaskiyu - “to put something on an object” (E.g. “Please put it on the chair”)
Verb takes 3 arguments, v(Agt, Pt, Obl.LOC)
Agent is typically ERG Pt is typically Abs Obl.LOC can be incorporated.
Surma dūy wasaskizekyašennaru. Surma dūy wasaski-zekyaš-enn-ar-u PN salmon put-table-3S.ANIM.ACC+3MS.ERG-PST-TRN “He put the salmon on the table.”
If this were Ainu, the incorporated Locative noun would license an Applicative affix; in Mohawk (and presumably its closely related Northern Iroquoian relatives), I haven’t seen incorporation of an oblique, only Patients. An exception is found in Tuscarora (Mithun p. 201 “Word Orders”, example 12a, contrasting non-incorporated 12b) in which we see a Locative arg incorporated in the stative verb Chukchi incorporates obliques, but often the theta role must be inferred by context. See A Lexical Account of Noun Incorporation in Chukchi (Silke Lambert, p.56). Minhast, as demonstrated in this article, incorporates obliques like Chukchi, but is much more restricted based on the semantic characteristics of the verb, i.e. verbs that incorporate obliques are restricted to a limited set of theta roles, and often the number of roles is restricted to one.
The semantic properties of a verb, namely the type of semantic/thematic role of an NP it can accept as an argument, influence what nouns may be incorporated. This semantic property limits the scope of which oblique nouns can incorporate, a limitation not exhibited in Chukchi. Motion verbs typically incorporate Goal or Origin nouns; Positional verbs typically incorporate Locative nouns. Otherwise, transitive verbs may incorporate either Patient or Instrument nouns; this sort of incorporation may seem to create ambiguities, but such is not the case, as both the polypersonal agreement pronominal affixes, and the presence or absence of an Instrumental Applicative make clear whether the Absolutive is a Patient or is an Instrument that has been promoted to the Absolutive argument. A few examples:
Example of a Goal argument: Iknatumankaran.
Ikna-tūman-ek-ar-an go-house-1S.NOM-PST-INTR “I went to/towards the house.”
Example of an Origin argument: Hahurtaħran.
ha-hūr-tah-ar-an come-mountain-2S.NOM-PST-INTR “You came from the mountain.”
Example of a Locative argument: Sap puħtabanakkaran.
Sap puħta-banak-ek-ar-an This stand-rock-1S.NOM-PST-INTR “I stood on this rock.”
Example of a Patient ABS with incorporated Instrumental argument: Redadesap ušnišuhapnekarun.
Redad=sap ušn-šuhapna-ek-ar-un Man this strike-sword-3MS.ACC+1S.NOM-PST-TRNS “I struck the man with this sword.”
Example of an Instrument ABS with incorporated Patient argument: Sapim šuhapna matušnerdattirkarun.
Sap min šuhapna mat-ušn-redad-tirk-ar-un This CONN sword INST.APPL-strike-3NS.ACC+1S.NOM-PST-TRNS “With this sword I struck the man.”
Noun incorporation is often associated with clauses that are structurally transitive, regardless of whether or not an oblique argument has been promoted to a core argument. However, in Minhast some stative verbs can noun incorporate. In other words, under certain circumstances, a clause that is structurally intransitive may also undergo noun incorporation. Such intransitive clauses tend to be nouns whose single core argument’s theta-role is that of Experiencer, e.g.
6) Kuldantuhamaran.
kuldan-tuham-ar-an sick-fever-3S.NOM.PST-INTR “He is sick with fever/He is sick and feverish.”
7) šuhapna wastanxundēban.
šuhapna wastan-xunde-ab-an Sword bleed-wound-3S.NOM.IMPF-INTR “He is bleeding from his sword-wounds/He is bleeding and wounded by sword/Because of that sword he was bleeding and wounded.”
8) Saxtisuspaħtayattaran.
saxt-suspaħ-tayatta-ar-an INCH-be.blind-poison-3S.NOM.PST-INTR “He became blind because of the poison.”
As an additional observation, those stative verbs that can incorporate tend to indicate sickness, injury, or congenital or other physical defects (as in Ex. #8 above). When the Inchoative prefix -saxt- occurs with the incorporated noun, native speakers tend to indicate the IN is the direct cause of the Experiencer’s state (Ex. #8). Otherwise the IN provides further details of or delimits the Experiencer’s current state (Examples #6 & #7), hence the alternative translations using the conjunction “and”. But this is not always so, as illustrated in the case of Ex. #7 which lacks the Inchoative. Here the sentence without -saxt- would seem to suggest that something else may have caused the bleeding, but there were other wounds that resulted from a sword. However, context and general knowledge that swords tend to cause bleeding would make that interpretation dubious. The use of the Inchoative -saxt- would definitely dispel that ambiguity, but is not necessary if context is sufficient to disambiguate between the two possible interpretations.
Also note in Ex. #7 the stranded NP “šuhapna” (sword): this stranded NP is the modifier of the incorporated noun “xunde”; the equivalent non-noun incorporated sentence would be “šuhapna min xundeyār wastanaban” (lit: From sword-wounds, he was bleeding), where the Ablative clitic =yār indicates the cause of the subject’s bleeding. The polypersonal agreement affix in Ex. #7 indicates there is only one core argument only, which would be the logical subject. This clearly demonstrates that “šuhapna” is a stranded NP.
Noun Incorporation in Intransitive Verbs
Although noun incorporation in Minhast is prototypically associated with lexically transitive verbs, lexically intransitive verbs may also incorporate nouns. The processes by which NI in intransitive verbs can be divided into three major categories:
- Nouns of non-Patient arguments (thematic relations)
- Body Parts
- Meteorological and other natural phenomena
Syntactic restrictions and other behaviours in intransitive NI otherwise that are extremely rare in transitive NI become more salient. Intransitive NI, where certain forms of irregularities occur, also reveal clues about the Proto-Nahenic ancestor that is more difficult to discern from transitive NI, as intransitive NI was more resistant to morphological leveling than transitive NI. This resistance can be attributed to the semantic features of the incorporated noun in the three aforementioned categories of intransitive NI.
Thematic Relations
In addition to incorporating semantic patients, which is what most people think about when talking about noun incorporation, is that Minhast, can incorporate a wider range of non-patient arguments, i.e. nouns with thematic relations differing from that of a semantic patient, can be incorporated into the Minhast verb.
Incorporated nouns interact with the semantics of its incorporating verb. Take for example the word kallutūyekaran (kallut-dūy-ek-ar-an), lit. "eat-fish-I-in.the.past-be". The incorporated noun is a semantic patient, and this is the prototypical type of noun incorporation that is seen most often in Minhast texts. In this example, incorporation of the patient noun has lowered the valency of the clause, as one would expect, and is further confirmed by the surfacing of the intransitive suffix -an.
Incorporation of Direction and Goal
Compare this with the transitive clause dūy kallutekaru < dūy kallut-ek-ar-un, lit. "Fish eat-I-in.the.past-do", a transitive clause wherein the transitive suffix -u also surfaces. Now compare this with the verb complex iknatūmanaran (ikn-tūman-ar-an), lit. "go-home-in.the.past-be". This time a noun (tūman "home"), whose incorporated noun's thematic role is that of Direction/Goal, has been incorporated. I.e. this isn't a semantically direct object incorporated into the verb; instead, a syntactically oblique argument has been incorporated into the verb complex. The non-incorporated form would be Tūman=aran iknaran (house-towards.it he went), and here we can see due to the postclitic directional case (usually called dative case) clitic =aran marks tūman "house" as an oblique argument, as it is not a core, absolutive argument.
Incorporation of Source
Classificatory noun incorporation, i.e. Type IV NI, occurs when the name of an overt toponym appears:
- Kattek sap harraħketappekaru.
/kat:ek sap har:aħkɛtap:'ekaru/
kattek sap raħk-han-tappe-ek-ar-u
place.name this APPL.ABL-come-place-3MS.ACC+1S.NOM-PST-TRNS
I came from this place (called) Kattek.
Note that the Ablative Applicative has surfaced to promote the place, named "Kattek", from an ablative oblique argument, i.e. the Source, to a derived Absolutive. The incorporated noun tappe ("place, location"), has been incorporated to specify Kattek as a location. As Kattek is a proper noun, and proper nouns lie high on the animacy hierarchy, Type IV is allowed so that Kattek can be backgrounded and later retrieved as a core argument.
This example also illustrates how noun incorporation can trigger modifier stranding. In this example, the demonstrative sap has been stranded from its head tappe. The absence of the min Connective is indicative that stranding has occurred. Demonstratives, when acting as modifiers, require the min particle to join it to its NP head. As a stranded modifier, sap is treated as an adjunct, able to be placed anywhere in the clause, including in postverbal position as an anti-topic.
Incorporation of Location
Another example, where a noun whose thematic role is of a non-Patient role, but rather that of Location, can be incorporated by interaction with the semantics of a verb, is saššihurran (sašši-hūr-ar-an), lit. "sit-mountain-in.the.past-be". Its non-incorporated counterpart is hurki sašširan (hūr=ki sašši-ar-an), lit. "mountain=on sit-in.the.past-be". Notice that these verbs which are incorporating non-Patient nouns are usually locational, positional, or motion verbs. However, oblique noun incorporation in transitive verbs is also allowable, and the incorporated noun is usually an Instrument, e.g. Dūy kallustespirtirkaru "I ate the salmon with my hand(s)." The Patient argument dūy ("salmon"), retains its position as an independent noun phrase in the absolutive case, whilst the oblique argument sespir ("hand"), is an incorporated noun whose thematic role is that of Instrument. The valency of the sentence has not changed, as the transitive -u suffix is still retained.
Incorporation of Experiencers
Verbs that take as their core NP with the Experiencer thematic relation often incorporate oblique nominals, whose theta role of Source or Cause, to background them, thereby focusing on Experiencer. The following two examples, the first with no incorporation, and the second with incorporation of the oblique nominal tipr ("meat") are semantically equivalent. The difference between the non-incorporated and incorporated versions is one of discourse purpose.
In the first example, the speaker is explicitly adding information about the cause of his sickness, the meat (tipr), with the Ablative clitic =yār. Additionally, by explicitly mentioning the meat he is introducing new information, as it had not yet been introduced into the discourse:
1) Non-incorporated NP:
- Tipiryār saxtikuldekaran
tipr=yār saxt-kuld-ek-ar-an
flesh=from INCH-sick-1S.NOM-PST-INTR
I became sick due to (infected) meat.
In the second example, another speaker's focus is on being sick. What caused her to be sick is of secondary importance, and there is an underlying assumption that both the speaker and the listener already know about the meat in question, either by previous discourse or other means. In this case, earlier in the week, the speaker and her brother had cleaned the refrigerator after an extended power outage. Unwilling to let an expensive cut of beef to go to waste, she foolishly ate it and got sick. Therefore, she chose to background the meat by incorporating the noun into the verb:
2) Incorporated NP:
- Saxtikulittipirkaran
saxt-kuld-tipr-ek-ar-an
INCH-sick-flesh-1S.NOM-PST-INTR
I got sick from the (infected) meat.
Incorporation of Body Parts
Body parts are often incorporated in attributive verbs. The pronominal affix represents both the subject of the clause (i.e. the syntactic pivot), as well as the possessor of the incorporated body part:
- Purrakyār saxtaharrumpakkaran.
purrak=yār saxt-tahal-ruppamak-ek-ar-an
dye=ABL INCH-be.green-face-1S.NOM-PST-INTR
My face became green from the dye (lit. "I became green face-wise from the pigment")
Overt NP possessors are stranded when their possessum is incorporated. In the next example, the overt possessor redad ("man") has been stranded after its possessum ruppamak ("face") was incorporated:
- Redad purrakyār saxtaharrumpakkaran.
redad purrak=yār saxt-tahal-ruppamak-∅-ar-an
redad pigment=ABL INCH-be.green-face-3S.NOM-PST-INTR
The man's face became green from the dye (lit. "The man from the pigment became face-wise green")
Speakers consistently reject sentences such as the following. In this infelicitous sentence, the NP has full possessor-possessum marking and occurs outside the verb complex, while the material that made the speaker's face green has been incorporated as if it were Instrumental argument. The restriction appears to be semantic. Unlike other stative verbs, attributive verbs select for incorporation the affected noun, not the Source or Cause noun:
- *Ruppamaktirekt saxtahalpurrakmaharan.
ruppamak-tirek=de saxt-tahāl-purrak-mah-ar-an
face-3NS.NOM+1S.NOM=ERG INCH-be.green-pigment-3NS.INAN.NOM-PST-INTR
(Intended): My face became green from the dye.
The correct version follows:
- Purrakyār saxtaharrumpakekaran.
purrak=yār saxt-tahal-ruppamak-ek-ar-an
pigment=ABL INCH-be.green-face-1S.NOM-PST-INTR
My face became green from the dye (lit. "I became green-faced from the pigment.")
Interaction with Applicatives
Interestingly, applicatives can be added to incorporating stative verbs to promote the oblique nominal whose thematic relation is that of Cause or Source. The Ablative Applicative -raħk- is usually selected for this operation. When the applicative is applied, the oblique nominal argument purrak is promoted to Absolutive status, thereby increasing the valency; nevertheless, the verb remains intransitive:
- Purrak saxtiraħkitaharrumpakekaran.
purrak saxt-raħk-tahal-ruppamak-ek-ar-an
pigment=ABS INCH-APPL.ABL-be.green-face-1S.NOM-PST-INTR
My face became green from the dye (lit. "Because of the pigment, I became green-faced.")
Incorporation of Meteorological Events and Other Natural Phenomena
Some natural phenomenon, like meteorological events, can undergo noun incorporation, as is the case with yam "sea", in this example:
- Intarsaššupninnaggammantassaraššamaran.
ntar-saxt-šupn-naggammantassarat-yam-ar-an
INCIP-INCH-SIM-be.tsunami-sea-PST-INTR
The sea started to rise high until it became almost like a tsunami.
Valency and Agreement Marking Irregularities
It remains debatable as to whether stative verbs with incorporated meteorological nouns are monovalent or zero-valent. The following example lends support to a monovalent interpretation: an overt non-null pronominal agreement marker -i- indicates that an underlying third person inanimate plural absolutive argument exists and has undergone pro-drop:
- Intarwakkayyuštiran.
ntar-wakkay-iyuššit-i-ar-an
INCIP-incidentally.meet-storm.cloud-3P.NEUT.ANIM.NOM-PST-INTR
Storm clouds gathered.
Polypersonal Marking
The primary purpose of the polypersonal markers in the Minhast verb are to reference the core arguments of its clause, whether they appear overtly, or are omitted through pro-drop. However, polypersonal marking can target the IN under certain restrictions. When agreement marking does appear, they tend to occur with collective or mass nouns, such as the inherently collective iyuššit. Moreover, this type of incorporation occurs with certain verbs only, particularly ones indicating movement, and even then iyuššit triggers agreement marking in a small fraction among these verbs.
- Nayyakiyuššitiyaran.
nayyaki-iyuššit-i-ar-an
gather.together-storm.cloud-3.ANIM.P-PST-INTR
Storm clouds gathered.
Oftentimes irregularities in gender-number concord may appear. An example where agreement marking is lacking with the very same collective noun follows in the next example. Note that the verb -adu- ("be many"), unlike -nayyaki-, is not a motion verb. The lack of an overt agreement marker with -adu- points towards a zero-valent interpretation:
- Intasaxtaduyyuštaran.
nta-saxt-adu-iyuššit-ar-an
INT-INCH-be.many-storm.cloud-PST-INTR
The sky became thick with storm clouds (lit. "The storm clouds became quite many.")
Similarly, when iyuššit is incorporated into the positional verb -puħt- "to stand upright", agreement marking does not occur:
- Puħtuyyuštaran.
puħt-iyuššit-ar-an
stand.upright-storm.cloud-PST-INTR
The storm clouds hung (over us) (lit. "The storm clouds stood upright")
Interestingly, -puht- can license agreement with other nouns, such as kayyūn "tree", when a collective meaning is intended:
- Yaššapuħtakayyummaharan.
yašša-puħt-kayyūn-mah-ar-an
there.DIST-stand.upright-tree-3S.NEUT-PST-INTR
The trees stood there.
Diachronic factors may explain the irregularities involving agreement marking for a subset of incorporated nouns interacting with a subset of verbs. The Proto-Nahenic ancestor originally had an extensive hierarchical noun class system, remnants of which remain in Minhast's relative Nahónda as evidenced by even more irregularities in the latter, and in its other relative Nankôre, whose elaborate nominal hierarchy may be a preservation of the protolanguage's original noun class system or an extensive elaboration of it. The irregular agreement marking triggered by -iyuššit- among a subset of a select class of verbs suggests that the noun once fell within a noun class of a particular animacy level. When the protolanguage split, the original noun class system were restructured in the daughter languages; further reductions and loss, particularly in both Minhast and Nahónda, left a residue in the form of the irregular agreement marking seen today.