UNIVERSIDAD DE SONORA
Maestría en Lingüística División de Humanidades y Bellas Artes
” Departamento de Letras y Lingüística
Cuerpos académicos “Estudios lingüístico-tipológicos y etnoculturales en lenguas indígenas y minoritarias” Universidad de Sonora (CA-81 )
īz in y “Análisis y documentación en lenguas indígenas” Escuela Nacional de Antropología e Historia (CA-14 )
aic innehnemiliz ye tlapalihcuiliuhtoc
ēnyo pohpolihu
In inihyo in int
“Su fama y su honra nunca perecerán, su historia en pinturas escrita está
Seminario decomplejidad sintáctica 2014 Seminario de complejidad sintáctica 2014
1 PROGRAMA
3 Sala de Usos Múltiples Departamento de Letras y Lingüística Edifício 3Q, 2o. Piso
Lunes 10 de noviembre
9:00-9:55 Nominalization, de-subordination and re-finitization T. Givón University of Oregon and White Cloud Ranch, Ignacio, Colorado
10:00-10:25 RECESO
10:30-10:55 Inflecting compounding in Mixe languages Rodrigo Romero Méndez IIFL-Universidad Nacional Autónoma de México
11:00-11:25 Topicalización en lacandón del sur Israel Martínez Corripio Escuela Nacional de Antropología e Historia
11:30-11:55 RECESO
12:00-12:25 The rise of the nominalizations: The case of the grammaticalization of clause types in Ecuadorian Siona Martine Bruil University of California at Berkeley
5 12:30-12:55 Adverbial clauses in Southeastern Tepehuan (O’dam) Gabriela García Salido Universidad de Sonora-CONACyT
13:00-15:55 COMIDA
16:00-16:55 On being an adjunct. Evidence from Algonquian Fernando Zúñiga University of Bern
17:00-17:25 RECESO
17:30-17:55 Posesión externa en mazahua Armando Mora-Bustos Universidad Autónoma Metropolitana-Unidad Iztapalapa
18:00-18:25 Lexical and clausal nominalization in Mochica Rita Eloranta Universiteit Leiden
18:30-18:55 The diachrony of grammatical nominalizations in Cahita (Uto-Aztecan) Albert Alvarez Gonzalez Universidad de Sonora
19:00 CENA LIBRE
6 Martes 11 de noviembre
9:00-9:55 On the role of person marking in finiteness and discourse Walter Bisang University of Mainz
10:00-10:25 RECESO
10:30-10:55 Temporal sentences in Yaqui: Topical arguments, coreference and switch-reference Lilián Guerrero IIFL-Universidad Nacional Autónoma de México
11:00-11:25 Destinative construction in Q’anjob’al (Maya): A complex predicate analysis Eladio Mateo Toledo (B’alam) Centro de Investigaciones y Estudios Superio- res en Antropología Social-Sureste
11:30-11:55 RECESO
12:00-12:25 Los predicados no finitos como construcciones desiderativas, y su restricción aspectual y de persona en el nawat de Pajapan, Veracruz Valentín Peralta Ramírez Escuela Nacional de Antropología e Historia
12:30-12:55 Clause chaining and nominalization in Tarahumara: A corpus oriented research Zarina Estrada Fernández y Jesús Villalpando Quiñónez Universidad de Sonora y University of Colora- do, Boulder
7 13:00-15:55 COMIDA
SECCIÓN DE POSTERS
16:00-16:20 Oraciones adverbiales temporales y correferencia de sujetos Rebeca Gerardo Tavira Maestría en Lingüística Hispánica Universidad Nacional Autónoma de México
16:20-16:40 La alternancia de a y para en construcciones complejas con verbos de movimiento: ¿oraciones finales o de propósito? Paola Gutiérrez y Valeria Benítez Maestría en Lingüística Hispánica y Doctorado en Lingüística-Universidad Nacional Autóno- ma de México
16:40-17:00 Comportamiento del verbo sentir en oraciones complejas: un estudio a partir de corpus Irasema Cruz Domínguez Maestría en Lingüística Hispánica Universidad Nacional Autónoma de México
17:00-17:25 RECESO
17:30-17:55 Entre pedir y el deseo. Gramaticalización del desiderativo en maayat’aan Fidencio Briceño Chel Centro INAH; Yucatán
8 18:00-18:25 The syncretism between antipassive and causative in Mocovi Cristian Juárez y Albert Alvarez Gonzalez Universidad de Sonora
19:00 CENA
9 RESÚMENES
11 Nominalization, de-subordination and re-finitization T. Givón University of Oregon and White Cloud Ranch, Ignacio, Colorado [email protected]
Across the Uto-Aztecan family, one finds a sharp distinction between the extreme nominalizing north and the resolutely finite south. The two northern-most sub-families, Numic and Takic, nominalize every subordinate clause in sight. The same seems to be true of Yaqui and Huichol. But further south one finds uniformly finite subordinate clauses in Tepiman (Tepehuan, Pima Bajo) all the way to Nahuatl. The transition zone between the two extre- mes, Guarijío, Trahumara and perhaps Cora, is surprisingly thin. There are good reasons for suggesting that the Uto-Aztecan north is both culturally (hunting-gathering) and linguistically (OV syntax) more conservative, and that the family’s south, due to either natural drift or contact with the Meso-American substratum is more innovative. So the question I would like to pose here is this: How does the drift from nominalized to finite subordinate clauses take place? Especially natural drift that is not induced by contact. The strategy I will pursue here is two-fold:
• Try to understand the internal logic of nominalization and related processes. • Try to find evidence, in the middle zone of the family and elsewhere, for the dynamics of change.
13 Inflecting compounding in Mixe languages Rodrigo Romero Méndez Seminario de Lenguas Indígenas, IIFl, UNAM [email protected]
Verb-verb compounding in Mixe-Zoque languages has been documented as one of the polysynthetic characteristics of these languages. It has been called nuclear serialization (Bril 2007; Foley & Olson 1985; Foley 1991; Romero 2009; Zavala 2000) in reference to the type of juncture (Foley & van Valin 1984). Thus, verb-verb compounding is a mechanism to create complex predicates, as shown in (1):
(1)Nyijkxy muum, jakam pues este... To’okëtejtp. y-nëjkx-y muum jakam pues este to’ok-jëtet-p 3S-go-DEP some.where far.away hmm DISC [3S]sell-walk-INDEP ‘He went far away hmm... He went to sell.’
On the other hand, compounding is also regarded as a morphological process usually restricted to word formation (cf. Lieber & Štekauer 2009), but, as Beck (2011) points out, there is no reason why we should confine, a priori, any formal mechanism to a particular function. According to this author, in Upper Necaxa Totonac compounding is also used to perform inflecting functions (something that Mel’cuk 2006:124 calls quasi-inflection), in addition to being used in word formation. The same situation occurs in Mixe: two roots can create a new lexical item, as in (2a), but compounding is also used to express grammatical meanings, as in (2b).
(2) a. Jëts, ¿pën käjpxtä’kp? jëts pëën käjpx-tä’äk-p and who speak-embroider-INDEP ‘And, who was praying?’
14 b. Pedro të pyujttsoony. Pedro të y-put-tsoon-y Pedro PERF 3S-run-go.away-DEP ‘Pedro started running.’
In both examples in (2), as well as in the example in (1), the last root bears the stress, something that affixes cannot do. In addition, the last root in these examples undergoes apophony (Ro- mero 2009), something unique to verb roots. Furthermore, this pattern is also found in nouns, as shown in (3). Thus, it is a widespread formal mechanism in the language use to convey grammatical meanings usually associated with inflection.
(3) tsäkäj-anä’äk cow-lads ‘cattle’
As Beck (2011) points out, it is possible that this situation exists not just Totonac, but in other polysynthetic languages, but it has previously missed. This paper addresses this issue.
References
Beck, David. 2011. Lexical, quasi-inflectional, and inflectional compounding in Upper Necaxa Totonac. In A. Aikhenvald & P. Muysken (eds.), Multi-verb Constructions: A view from the Americas, 63-106. Leiden: Brill. Bril, Isabelle. 2007. Nexus and juncture types of complex predicates in oceanic languages: Functions and semantics. Language and linguistics 8 (1): 276-310. Foley, William, & Olson, Mike. 1985. Clausehood and verb serialization. In J. Nichols and A. C. Woodbury (eds.), Grammar inside and outside the clause: Some approaches
15 to theory from the field, 17-67. Cambridge, UK: Cambridge University Press. Foley, William, & Van Valin, Robert D. Jr. 1984. Functional syntax and universal grammar. Cambridge: Cambridge University Press. Foley, William. 1991. The Yimas language of New Guinea. Stanford: Stanford University Press. Lieber, Rochelle & Pavol Štekauer. 2009. The Oxford handbook of compounding. Oxford, UK: OUP. Mel’cuk, Igor. 2006. Aspects of the theory of morphology. Berlin: Mouton de Gruyter. Romero Méndez, Rodrigo. 2009. A reference grammar of Ayutla Mixe. University at Buffalo: Ph.D. Dissertation. Zavala Maldonado, Roberto. 2000. Inversion and other topics in the grammar of Olutec (Mixe), University of Oregon: Ph.D. Dissertation.
Topicalización en lacandón del sur Israel Martínez Corripio Escuela Nacional de Antropología e Historia [email protected]
En este trabajo presento algunos rasgos respecto a la topicaliza- ción en lacandón del sur, lengua maya de la rama yucatecana. La topicalización se utiliza para atraer la atención del oyente hacía un participante en el discurso, de tal forma que el constituyente topicalizado regularmente es definido y, preferiblemente, se colo- ca frente al predicado (Aissen 1992: 50). En las lenguas mayas yucatecanas, el marcador de tópico -e’ sirve para indicar límites de cláusulas que incluyen cláusulas relativas, condicionales y su-
16 bordinadas adverbiales (Bohnemeyer 1998; Hofling 2000; Bergqvist 2008). En lacandón del sur, específicamente, el elemento léxico topicalizado puede aparecer antes o después del predicado como se muestra en (1a) y (1b). Por lo tanto, al igual que en otras len- guas yucatecanas, el lacandón del sur distingue al elemento topicalizado por medio de la marca -e’, que la mayoría de las veces tiene correlación con el determinante a. Sin embargo, exis- ten casos en los cuales el topicalizador aparece sin este determi- nante (1c).1
(1)a. peero a baarum-e’ k-u-ka’a=nuk-ik-ø pero DET jaguar-TOP IPF-A3-NUM=contestar-ICTR-B3 ‘pero es el tigre el que vuelve a contestar’ (HombreTigreJCHK014)
b. ba’k a-kuch-(i)k-ø a teech-e’ qué A2-cargar-ICTR-B3 DET EF2-TOP ‘entonces ¿qué cargas tú?’ (JNSQC059)
c. chen ts’om-e’ k-u-mäk-ik-ø sólo seso-TOP IPF-A3-comer-ICTR-B3 ‘sólo sesos come’ (HombreMono04)
El rol gramatical del constituyente topicalizado puede ser el de sujeto de una estructura transitiva, sujeto de una estructura
1 Abreviaturas: A3 Ergativo Tercera Persona A2 Ergativo Segunda Perso- na B3 Absolutivo Tercera Persona DET Determinante DI Deíctico EF2 Enfático Segunda Persona ICTR Incompletivo Transitivo IPF Imprerfectivo NUM Numeral TOP Topicalizador.
17 intransitiva, sujeto de un predicado no verbal o bien, el objeto de una estructura transitiva. Por otro lado, los elementos que no son argumentos del núcleo también pueden ser topicalizados, lo cual ya ha sido reportado para otras lenguas mayas como el chol (Vázquez-Álvarez 2011) y el tseltal (Polian 2013). De tal forma que en lacandón del sur los adverbios, los locativos y cláusulas completas, se pueden topicalizar a partir de -e’. Finalmente, este marcador también se utiliza para indicar los límites de diferentes tipos de cláusulas, como cláusulas relativas, condicionales y cláu- sulas adverbiales subordinadas. En este trabajo presento con ma- yor detalle cada uno de estos casos.
Referencias
Aissen, Judith. 1992. “Topic and Focus in Mayan”, en Language 68, 43-80. Bergqvist, Henrik G. 2008. Temporal Reference in Lakandon Maya: Speaker-and Event-perspectives. PhD. Dissertation. University of London. Bohnemeyer, J. Bohnemeyer, Jürgen. 1998. Time relations in discourse: Evidence from a comparative approach to Yukatek Maya. PhD. Dissertation. Tilburg University. Hofling, Charles A. 2000. Itzaj Maya Grammar. The University of Utah Press. Polian, Gilles. 2013. Gramática del tseltal de Oxchuc. Centro de Investigaciones y Estudios Superiores en Antropología So- cial. México. Vázquez Álvarez, Juan Jesús. 2011. A Grammar of Chol, A Mayan Language. PhD. Dissertation. University of Texas at Austin.
18 The rise of the nominalizations: The case of the grammaticalization of clause types in Ecuadorian Siona Martine Bruil University of California at Berkeley [email protected]
Ecuadorian Siona, a Western Tukanoan language spoken in the eastern lowlands of Ecuador, marks the clause type of a sentence or clause in its verbal morphology. The language has distinct marking for assertive, interrogative, reportative and dependent clauses. This marking of clause type consists of a portmanteau suffix that also marks subject agreement and tense, and some additional clause type specific morphology. Interestingly, the portmanteau subject agreement suffixes for interrogative, reportative and dependent clauses are very similar in form to the nominalizers in the language. The similarities in form are illustrated in the examples below:
(1) a. sa-i-hi. (Assertive) go-IMPF-3S.M.PRS.ASS ‘He is going.’
b. sa-i-ki? (Interrogative) go-IMPF-2/3S.M.PRS.N.ASS ‘Are you (M)/ is he going?’
c. sa-i-ki-jã. (Reportative) go-IMPF-2/3S.M.PRS.N.ASS-REP ‘You (M) are/ he is going, they say.’
d. sa-i-ki-na jã-wi. (Dependent) go-IMPF-S.M.PRS.DEP-DS see-OTH.PST.ASS ‘While he was going, I saw (him).’
19 e. sa-i-ki-bi jã-bi. (Nominalization) go-IMPF-NLZ.S.M-SBJ see-3S.M.PST.ASS ‘The one who was going saw (it).’
The assertive suffix -hi in (1a) is different from the other clause- typing portmanteau suffixes. The form -ki is used for interrogative (1b), reportative (1c), and dependent clauses (1d). Additionally, it is also the masculine agentive nominalizer (1e). The difference between these clause types and the nominalization is marked by means of additional morphology, such as the reportative suffix -jã (1c), the dependent different subject suffix -na (1d), and the subject case marker -bi (1e). Another way to differentiate between the clause types is the organization of the subject agreement categories. For instance, in interrogative and the reportative clauses -ki marks only second and third person singular masculine, while in dependent clauses it marks all singular masculine persons. In this paper, I will claim that the similarities between the markings of the different clause types are not a coincidence. I will propose that nominalizers were used in the past to mark complement clauses that were used in interrogative and reportative constructions and in adverbial subordination. These constructions grammaticalized as the clause- typing system that is found in Ecuadorian Siona today.
Adverbial clauses in Southeastern Tepehuan (O’dam) Gabriela García Salido Universidad de Sonora-CONACyT [email protected]
Strategies for adverbial clauses in Southeastern Tepehuan include subordinating morphemes and word order. Furthermore, the language has morphology that is exclusively used for adverbial
20 clauses (1), as well as adverbial morphology that is shared with other subordinating structures, such as complement clauses and relative clauses (2). With respect to adverbial-clause marking (e.g., affixes, particles, and order), Southeastern Tepehuan exhibits a unique behavior insofar as second-position clitics act as an indicator of thematic continuity for the subject (1-2). This suggests that these clitics have evolved independently with the function of marking switch reference. This presentation, based on a corpus of 30 hours of data, sheds light on formal and semantic resources for encoding adverbial clauses, as well as the communicative functions of adverbial clauses in discourse. It further shows how this language conveys adverbial meanings by non-embedded clauses (as with juxtaposition for ‘when’ clauses), as in (3), and possibly in a later stage by embedded clauses introduced by different subordinators (i.e., overt morphology (1-2)), in a similar way to creole and Australian languages (Cristofaro 2003). This talk will also allow us to understand the complexity of O’dam and the way speakers establish coherence with respect to several aspects of its grammar, such as syntactic dependency; the roles of the participants; tense, aspect, and mood markers; as well as the communicative function of these clauses.
(1)Px pui’ muk-ix-kam mu kat MIR SENS die-RES-POSP:origin DIR lie.down na=ñich-pai’dhuk mu jii SUB=1SG.SBJ.PFV-ADVR DIR go.PFV ‘He was already dead, lying down/on the ground when I went there.’ (Text_102010_HMA_GGS_Suesposo, 03:38)
21 (2)Mu sap na=t ai ba-chia-mt DIR REP.UI SUB=3SG.SBJ.PFV arrive CMP-order-3PL.SBJ.PFV na tbañ-dha’ ma’n kaja gu koka SUB take.down-APPL one box DET cokes ‘When he arrived there, they asked him to take down a box of cokes.’ (Text_102010_PSC_GGS_Lavidademiesposo, 30:13)
(3)Ba-ji-chu sas-ji gui’ sap bhammk CMP-start-CAUS play-DC DEM REP.UI DIR leave.PFV
‘Hei started playing when hei left.’ (Text_092011_MCC_GGS_Elhielo, 07:49)
References
Comrie, B. 1983. Switch-Reference in Huichol: A Typological Study. Switch-Reference and Universal grammar. Typological Studies in Language 2, ed. by J. Haiman and P. Munro, 17-37. Amsterdam: John Benjamins Publications. Cristofaro, S. 2003. Subordination. Oxford; New York: Oxford University Press. Mithun, M. 1993. Switch Reference: Clause Linking in Central Pomo. International Journal of American Linguistics 59, 119-137.
22 On being an adjunct. Evidence from Algonquian Fernando Zúñiga University of Bern [email protected]
The Algonquian languages of North America are well known for phenomena like animacy-based gender, obviation, and morphosyntactic inversion, as well as for the intricate verb morphology they show. The issue of grammatical relations, however, in particular the question of how Algonquian syntactic functions are to be compared with those developed based on fa- miliar western European languages, has received attention only comparatively recently. Work on Ojibwe (Central Algonquian) by Rhodes (2006, 2010) proposed an Algonquian-specific grammatical relation called “relative root complement” (RRC) that came to complement the inventory of functions to be acknowledged descriptively and theoretically. The present paper argues that, in Blackfoot (Western Algonquian), RRCs are not distinct from secondary objects, and that the most striking feature in the syntax of the language is probably the fact that run-of-the-mill adjuncts may not be found at all. This leaves us with a twofold intriguing picture to be further explored by in-depth research: (i) that Blackfoot has, unlike Ojibwe, not actually more grammatical relations than non-Algonquian languages, but less, and (ii) that adjuncts may not be universal at all.
23 Posesión externa en mazahua Armando Mora-Bustos Universidad Autónoma Metropolitana-Unidad Iztapalapa [email protected]
El objetivo de esta presentación es describir y caracterizar las construcciones de posesión externa en mazahua (lengua del subgrupo pame perteneciente a la familia otomangue) y establecer su pertinencia dentro del ámbito translingüístico. En esta lengua las frases nominales posesivas que aparecen en función de objeto di- recto expresan una relación de posesión entre el poseedor y lo poseído a través de una forma de posesión interna o de una forma de posesión externa (Payne y Barshi 1999), como en (1). En la frase nominal compleja de (1a), ot’i i nuzo ‘el niño del señor’, se establece una relación de posesión interna. La FN nu zo modifica al sustantivo t’ii . Esta relación está marcada por el prefi- jo posesivo o- . Por su parte, en (1b), la relación de posesión ex- terna que se establece entre el poseedor nu zo ‘señor’ y lo poseído ot’i i ‘el niño’, se codifica por fuera de la FN, en un mor- fema de dativo en el verbo. El dativo hace referencia cruzada con el poseedor.
(1)a. o-ma O-got’ö=hi o-t’ iiÛ nu= zo 3.PST-ir 3.PST-encerrar=PL 3POSS-niño ART=señor ‘encerraron al niño del señor’
b. o-ma O-gor-p’ö=hi o-t’ iiÛ nu= zo 3.PST-ir 3.PST-encerrar-3DAT=PL 3POSS-niño ART=señor ‘le encerraron al niño del señor’
En (1b) aparece una construcción de posesión externa con dativo. En este tipo de construcción la unidad gramatical que re- fiere al poseedor funciona como un típico objeto indirecto. Tanto
24 el objeto indirecto como la FN que refiere al poseedor aparecen después del objeto directo (SVO-OI); igualmente, el objeto indi- recto como el dativo se codifica en un repertorio amplio de sufijos verbales que hacen referencia cruzada con una FN. Es imperativo conocer la distribución de la construcción de posesión externa; Haspelmath (1999) estable un conjunto de je- rarquías para la construcción de posesión externa de las lenguas indoeuropeas. De este conjunto de propiedades (animacidad del poseedor, situación, inalienabilidad de lo poseído y relación gra- matical de lo poseído); aquí, se mostrará la relevancia y la jerar- quía, como en (2). Generalmente, el poseedor es una entidad ani- mada, el estado de cosas que expresa el verbo corresponde a una dinámica con afectación, la entidad poseída es inalienable y la re- lación gramatical que expresa lo poseído es el objeto directo.
(2)a. riÛ-phبr-k’ö i-t iiÛ 1.FUT-cuidar-2DAT 2POSS-niño ‘cuidaré a tus niños’
b. nu=Pedro o-j’ بkö-zö i-ñiÛ i-j’Û =khØ ART=Pedro 3.PST-cortar-1DAT 1POSS-cabeza-mano=1POSS.ENF ‘Pedro me cortó el dedo’
c. nu=nda m8a o-pa ¨ -giÛ i-t hus’ö ART=viento 3.PST-aventar-1DAT 1.POSS-sombrero ‘el viento aventó mi sombrero’
A lo largo del trabajo se presentará una variedad de datos que apoyen la caracterización, distribución y restricción de este tipo de construcción. Tipológicamente estas construcciones de dativo son propias de las lenguas indoeuropeas; en las lenguas originarias de México han sido identificadas en el otomí (Palancar 2009), en consecuencia, es importante resaltar el hecho de que estas cons- trucciones son viables en otros sistemas.
25 Referencias
Aikhenvald Alexandra Y. y R.M.W. Dixon. 2013. Possession and ownership. A cross-Linguistic Typology. Oxford. Oxford University Press. Haspelmath, Martin. 1999. ‘External Possession in a European Areal Perspective’ en External possession. Doris L. Payne y Immanuel Barshi (eds). Amsterdam. John Benjamins Publishing. Ortiz Alejandra, Armando Mora-Bustos, Francisco Arellanes y H. Antonio García. 2014. ‘Codificación de las relaciones semánticas no eventivas en cuatro lenguas otomangues’. VI Coloquio sobre lenguas Otomangues y Vecinas Mario Molina Cruz. Oaxaca. Oaxaca. Palancar Enrique L. 2009. Gramática y textos del hñoñho . Otomí de San Ildefonso. Tultepec, Querétaro. México. Universi- dad de Querétaro y Plaza y Valdes. Payne, Doris L. y Immanuel Barshi. (eds.). 1999. External possession. Amsterdam. John Benjamins Publishing.
Lexical and clausal nominalization in Mochica Rita Eloranta Universiteit Leiden [email protected]
Mochica is an extinct linguistic isolate spoken until the mid-to-late nineteenth century in the northern coastal area of Peru. Mochica is typologically and different than other Andean languages. The Mochica language has been preserved whilst it was still been spoken in two colonial documents: Rituale, Sev Manvale Pervanvm by Jerónimo de Oré (1607) and Arte de la lengua
26 yunga, a grammatical description by Fernando de la Carrera (1644). The latter is the main source for the data analyzed in the present investigation. Despite nominalization being prominent in Mochica there has not been systematic treatment of Mochica nominalization pro- cesses so far. In this presentation I aim to offer a complete classification and description of Mochica lexical and clausal (grammatical) nominalizations. The subordinate clauses in Mochica are formed by means of nominalizing and subordinating suffixes. Subordination is not only restricted to complement clauses (1), but concerns also relative (2) and adverbial clauses (3) as shown in the following examples:
(1)Confessar læ-çæc fe p-oc penitençia. confess be.NMLZ.RESULT COP PASS-name/call penance ‘The confession is named/called penance’.
(2)Tzhang ai-apæco tzhang chi-co-pæco Dios 2SG do-NMLZ.AG 2SG be-CAUS-NMLZ.AG God ‘God who has made and created you’
(3)Mæiñ ef ang læm-ædo tzhang Lima-c 1S.GEN father 3S COP die-PART 2SG Lima-LOC chilæc TO BE-NMLZ ‘My father died when you were in Lima’
Besides these cases of clausal nominalizations, I will present some other nominalizers such as <-næm>, <–chæm> and <–top>. <–næm> (called dative gerund by De la Carrera 1644: 53, 60) is a purposive nominalizer and is used in connection with verbs meaning ‘to wish’ (
27 (4) Funo-næm eiñ-loc Eat NMLZ.PURP 1SG-wish/want ‘I want to eat’.
<–chæm> which was called “future in – rus” by De la Carrera (1644: 202) is a nominalizer that has, for instance, the modality of obligation, see example (5):
(5)Chi-ñ-chæm be-1SG-NMLZ.OBL
Finally, <–top> is a sequential suffix that expresses a sequence of events or actions as shown in (6):
(6)al-top olo infierno- ng nic çoc descend-NMLZ.SEQ fire hell- GEN LOC three lun-ær nico choc- top day-GEN LOC get up NMLZ.SEQ ‘(He) descended to the fire of hell and after three days (he) rose’
<-næm>, <–chæm> and <–top> have been previously considered subordinate suffixes (Hovdhaugen 2004: 46) but in this study I intend to demonstrate that they form part of the complex Mochica nominalizers system. The final intention of this research is to offer a systematic analysis of all processes of nominalization in Mochica.
References
Carrera, Fernando de la. 1644. El Arte de la Lengva Yvnga de los valles del Obispado de Truxillo del Peru, con vn Confessonario, y todas las Oraciones Christianas, tradu-
28 cidas en la lengua, y otras cosas, de Fernando de la Ca- rrera Daza, impreso en Lima. Lima: Joseph Contreras. Hovdhaugen, Even. 2004. Mochica (Languages of the World/ Materials 433). Munich: LINCOM. Oré, Luis Jerónimo. 1607. Rituale, Sev Manvale Pervanvm, Naples: Iacobum Carlinum, & Constantinum Vitalem.
The diachrony of grammatical nominalizations in Cahita (Uto-Aztecan) Albert Alvarez Gonzalez Universidad de Sonora [email protected]
Based on the approach developed by Shibatani (2009; also in Shibatani & Awhad 2009) concerning the connection between relativization and nominalization in different languages around the globe, in a previous work (Alvarez 2012) I have proposed that “relative clauses” in Yaqui (a language from the Uto-Aztecan family spoken in the Northwest of Mexico) are in fact grammatically nominalized expressions. Considering mostly synchronic evidences, I have argued that these grammatical nominalizations may have two major functions (referential and noun-modifying functions) depending on the syntactic uses of the nominalized expression (respectively, as NP-head and as NP-dependent in apposition), and that relativization in Yaqui has to be actually considered as merely one specialized function of nominalization: the modifying function of an appositive grammatical nominalization. In this contribution I will present new diachronic evidences in support of this nominalization approach. In doing so, I will survey the origin of the nominalization markers in Cahita (Yaqui, Mayo,
29 …Tehueco) as well as the evolution undergone by the nominalized expressions marked by the suffix -(’)u which has changed from being an old agent nominalizer with tempo-aspectual restrictions to being the current patient nominalizer. I will show that these diachronic data clearly advocate for the nominal status of the supposed “relative clauses” in Cahita.
References
Alvarez, A. 2012. Relativization and Nominalizations in Yaqui. In Relative clauses in Languages of the Americas. A typological overview. [Typological Studies in Language 102], Comrie, B. and Z. Estrada-Fernandez (eds.), 97-126. Amsterdam: John Benjamins. Shibatani, M. 2009. Elements of complex structures, where recursion isn’t it. The case of relativization. In Syntactic Complexity. Diachrony, acquisition, neuro-cognition, evolution [Typological Studies in Language 85], Givón, T. and M. Shibatani (eds), 163-198. Amsterdam: John Benjamins. Shibatani, M. and K. Awadh Bin Makhashen. 2009. Nominalization in Soqotri, a South Arabian language of Yemen. In Endangered languages: Contributions to Morphology and Morpho-syntax, L. W. Wetzels (ed.). Leiden: Brill. 9- 31.
30 On the role of person marking in finiteness and discourse Walter Bisang Universidad de Mainz [email protected]
Person marking is used for reference tracking within individual sentences as maximalsyntactic units as well as beyond sentences within units of discourse. In addition to this function and motivated by it, the same markers can also be used as indicators of finiteness at the level of the sentence and, in the case of co-subordination (Van Valin 2005), also at the level of discourse units (cf. Bisang 2007 on finiteness/non-finiteness asymmetries). In fact, it is often hard if not impossible to find criteria for clearly distinguishing sentences from discourse units in co-subordination. Reference-tracking can be expressed by elaborate systems of person marking as well as by zero marking (zero pronouns). The present paper will show that morphologically complex switch- reference systems with different sets of person marking for nonfinite/dependent and finite/independent verb forms (e.g. in Amele and Fore, Papua New Guinea) can serve the same purpose of creating discourse units as zero pronouns in East and mainland Southeast Asian languages like Chinese, Khmer or Thai. Two cases will be discussed. One case is the creation of discourse units by the use of finite verb morphology instead of switch-reference morphology. The other case is the use of different- subject markers even though there is no change of subject. This phenomenon is well-known for a number of switch-reference languages (e.g. Roberts 1988). Typical criteria are deictic continuity vs. deictic change in the area of time, place or world. If there is no deictic change, a marker from the same-subject system is selected, while the different-subject system is employed in instances of deictic change.
31 In both cases, East and mainland Southeast Asian languages use the opposition of overt person marking by a pronoun vs. zero marking. The similarity of the discourse functions expressed by these two rather different types of marking is quite remarkable. As I will argue in this paper, the two systems represent two rather extreme poles of two different types of complexity, i.e., overt vs. hidden complexity (Bisang 2009, 2014). More particularly, they can be seen as the result of two different types of maturation. The switch- reference system is a result of explicitness oriented maturation in the sense of Dahl (2004), while the systems based on zero vs. pronoun are instances of economy-oriented maturation.
References
Bisang, W. 2007. Categories that make finiteness: discreteness from a functional perspective and some of its repercussions. In: Nikolaeva, Irina (ed.), Finiteness. Theoretical and empirical foundations, 115-137. Oxford: Oxford University Press. Bisang, W. 2009. On the evolution of complexity-sometimes less is more in East and mainland Southeast Asia. In: Sampson, G.; Gil, D. & and Trudgill, P. (eds.), LanguageComplexity as an Evolving Variable, 34-49. Oxford: Oxford University Press. Bisang, W. 2014. Overt and hidden complexity-two types of complexity and their implications. Poznan Studies in Contemporary Linguistics 50(2), 2014, 127-143. Dahl, Ö. 2004. The Growth and Maintenance of Linguistic Complexity. Amsterdam/ Philadelphia: Benjamins. Roberts, J. R. 1988. Amele switch-reference and the theory of grammar. Linguistic Inquiry, 19.1, 45-63. Van Valin, R. D. Jr. 2005. Exploring the Syntax-Semantics In- terface. Cambridge: Cambridge University Press.
32 Temporal sentences in Yaqui: Topical arguments, coreference and switch-reference Lilián Guerrero IIFL-Universidad Nacional Autónoma de México [email protected]
This paper offers a corpus-based study of temporal adverbial clauses in Yaqui. Temporal clauses are introduced by two major adverbial subordinators: -kai (-ka when the clause is non-final), as in (1a), and -o, as in (1b); in addition, specific temporal clauses as those in (2) are also possible. In a previous work, I suggested that the analysis of Yaqui temporal clauses turns particularly problematic for the following reasons. First, the two most common linkage markers encode other adverbial relations, i.e., they are semantically ambiguous. Thus, simultaneous and sequential relations are not formally distinguished. Hence, one may expect that clauses like those in (2) are more frequent in discourse than clauses in (1). This initial hypothesis turns to be false. Since general temporal clauses marked by o-/-kai are the most frequent in corpus, one may wonder how posterior vs. anterior temporal relations are distinguished. Two valid hypotheses: (i) the ordering between the main and adverbial units, (ii) the TAM information on the verb. With respect to ordering, and regardless of the temporal relation between the two events, the adverbial unit tends to be sentence initial, meaning the sequential iconicity motivation is not relevant. We cannot fully rely on TAM information either since o-, and mainly kai-clauses, tend to be unmarked. In this paper, the analysis focuses on the argument realization and coreferential patterns found inside temporal clauses. On one hand, kai-clauses demand a missing syntactic argument in the adverbial unit, which must be in co-reference with the main subject. Since temporal clauses tend to occur sentence-initial, then the lexical subject is overtly expressed in the following clause, i.e., cataphoric
33 relations. There are a few cases where the subject is expressed as a nominative argument at the beginning of the sentence, i.e., nominative subjects are unexpected in subordinated clauses. In contrast, o-clauses prefer different-subjects, but still there are a few examples of identical subjects in the corpus. Regardless the identity of the subject, the dependent subject must be overt and coded accusative inside the adverbial unit, i.e., cataphoric relations are uncommon. In terms of argument coding, specific temporal clauses are less tight when compared to general temporal clauses: the two subjects in (2) are also identical, but there is a co-referential nominative pronoun in the main unit.
(1)a. [bea sechupti _i pensasaroa-ka]1º nei aman MD suddenly think-CLM 1SG.NOM there
siika2º go.SG.PFV ‘And, when I suddenly thought [about it], I went there.’
b. [Ju-ka Sulumai-tai omotria-u yepsa-k-o]1º DET-ACC Sulumai-ACC brush-DIR arrive-PFV-CLM
jaibu _i kaa enchi tea-k2º already NEG 2SG.ACC find-PFV ‘When Sulumai got back to the brushes, she couldn’t find you.’
(2)[kee Sulumai-tai bwij-wa-o]2º aapoi enchi ADV.NEG Sulumai-ACC capture-PAS-CLM 3SG.NOM 2SG.ACC
juya-m nasuk e’e-ria-k1º brush-PL middle hide-APPL-PFV ‘Before Sulamai was captured, she was able to hide you inside the brush.’
The marking of the adverbial clause, the coreferential patterns and the lexical coding of shared participants in Yaqui temporal
34 clauses will be discussed in terms of the well-known system of switch-reference marking.
Destinative construction in Q’anjob’al (Maya): A complex predicate analysis Eladio Mateo Toledo (B’alam) Centro de Investigaciones y Estudios Superiores en Antropología Social (CIESAS) [email protected]
Purpose constructions are complex clauses with two situations linked by a PURPOSIVE RELATION where the matrix predicate is performed with the intention/goal of obtaining the realization of a situation, the purpose clause (Schmidtke-Bode 2009: 20, and others). This means that they involve intentionality on the part of an argument of the main clause, the purpose clause has an intrinsic future orientation, and the outcome is intended/desired or hypothetical, like in (1). Q’anjob’al (a Mayan language spoken in Guatemala) has finite purpose clauses and motion-cum purpose clauses, example (2), whose properties are no different from those of the English purpose clause. However, Polian, Mateo and Can (2013), following Simonin’s (2011) analysis of weak purpose clauses in English, show that Q’anjob’al and Tseltal (Mayan) have a DESTINATIVE CONSTRUCTION defined as ‘the construction denotes a situation where the matrix verb makes available an entity that is earmarked for a particular use, specified by the second verb’. The destinative is shown in (3); it resembles purpose clauses, but differs from them in structure and semantics, and has been unrecognized in the typology of purpose clauses (Schmidtke-Bode 2009). One
35 difference is that intentionality is not necessary, (3). Polian, Mateo and Can suggest that destinative constructions are complex clauses that differ from infinitives and nonfinite clauses. The goal of this paper is twofold. First, I provide a detailed description of the features the destinative construction, which is not yet available, the data come from my own corpus of natural texts. Second, I show that destinative constructions share the properties of nonfinite clauses and complex predicates in Q’anjob’al. In particular, I show that the destinative construction with transitive verbs patterns like ditransitive and causative complex predicates in that they all involve argument structure fusion (Butt 1995; Mateo 2012; among others). The ditransitive complex predicate in (4) and the destinative in (3) have argument fusion; their second verbs lack inflection of their ABS2SG logical objects (vs. the nonfinite clause in (2)) and this cannot be explained under control. However, destinative constructions with intransitive verbs are syntactically similar to infinitives whose features can be explained under control, though they undergo integration that resembles resultative serial verbs.
(1) A: A monkey picked leaves or fruit to eat them. B: So did it eat them? A: I have no idea, but that was certainly its intention. {based on Simonin 2011:2}
(2) Kax max-ach b’et hin ha-kol-on-i then COM-ABS2SG go_return ABS1SG ERG2SG-help-DEP-FF ‘Then, you went [to help me].’
(3)chot-an hach ek’ j-il-a’ sitting-POSS ABS2SG DIR ERG1PL-see-TV ‘You are sitting for us to see you.’
36 (4) Ch-ach ul hin-say w-il-a’ INC-ABS 2SG come ERG1SG-look.for ‘I come to look for you (for myself).’ {txt062}
Butt, Miriam. 1995. The Structure of Complex Predicates. Stanford, CA: CSLI Publications. Mateo Toledo, Eladio. 2012b. Complex Predicates in Q’anjob’al (Maya): The Verbal Resultative. International Journal of American Linguistics vol. 78(4), pp. 465-95. Polian, Gilles, Eladio Mateo and Telma Can Pixabaj. 2013. Cons- trucciones destinativas en lenguas mayas. ms. Schmidtke-Bode, Karsten. 2009. A typology of purpose clauses. Amsterdam: John Benjamins. Simonin, Olivier. 2011. Adverbial and relative to-infinitives, Journal of English Linguistics, available online at: http:// eng.sagepub.com/content/early/2012/01/03/00754242 11428337.
Los predicados no finitos como construcciones desiderativas, y su restricción aspectual y de persona en el nawat de Pajapan, Veracruz Valentín Peralta Ramírez Escuela Nacional de Antropología e Historia. México [email protected]
El objetivo principal de este trabajo es describir las propiedades morfosintácticas de ciertas cláusulas de complemento que se pre- sentan en el nawat de Pajapan, una variedad del nawat que se habla en el sur del estado de Veracruz, México. De acuerdo a los
37 objetivos particulares de esta presentación, voy centrarme en aque- llas cláusulas complejas cuyas cláusulas de complemento están reducidas en sus marcas tiempo-aspectuales, manifestándose una mayor dependencia tanto sintáctica como semántica con la cláu- sula matriz, y por lo tanto, pueden ser consideradas como cláusu- las complejas con mayor integración clausal. Esto último, puede ser confirmado por el mecanismo de la incorporación, esto es, siendo el nawat una lengua polisintética (cf. Baker 1996) y con marcación en el núcleo (cf. Nichols 1986 ), los complementos oracionales no-finitos pueden ser incorporados al verbo matriz, confirmándose así una mayor integración entre la cláusula matriz y su complemento. Por lo tanto, en este trabajo voy a presentar tres tipos de cláusulas de complemento, las cuales serán consideradas como cláusulas complejas con una mayor dependencia con el pre- dicado matriz:
(1) ni-mis-go:wi-li:-s-negi mo-tegak 1SUJ-2OP-comprar-APL-IRR-querer 3POS-ZAPATO ‘Quiero comprarte tus zapatos.’
(2)ni-k-negi [ ni-mitz-go:wi-li:-s mo-tekak ] 1SUJ-3OP-querer 1SUJ-2OP-comprar-APL-IRR 2POS-zapato ‘Quiero comprar tus zapatos.’
Todas estas construcciones comparten, al menos dos rasgos similares, por ejemplo, el evento expresado por el verbo de com- plemento es posterior al evento expresado por el predicado ma- triz, y todas estas construcciones presentas restricciones de per- sona. Estos rasgos permiten pensar que este tipo de construcciones son más dependientes del predicado matriz, y presentan distintos grados de integración clausal, de modo que, como predicados no finitos son más cercanos a los predicados nominalizados. Estos factores sintácticos y semánticos presentes en la unión de cláusu- las complejas serán los tópicos centrales de esta presentación.
38 Refeencias
Baker, Mark C. 1996. The Polysynthesis Parameter. Oxford University Press, New York. pp. 338-396. Cristofaro, Sonia. 2003. Subordination. Oxford: Oxford University Press. Dixon, R. M. W. 1995. Complement clauses and complement strategies. Meaning and Grammar, F.R. Palmer (ed). Cambridge: Cambridge University Press, 174-220. Givón, T. 1980. The binding hierarchy and the typology of complements. Studies in Language 4.3, 333-377. Nichols, Johanna. 1986. “Head-marking and dependent-marking grammar”, en: Language 62. pp. 56-119. Noonan, M. 1985. Complementation. Language Typology and syntactic description. Volume 2.
Clause chaining and nominalization in Tarahumara: A corpus oriented research Zarina Estrada Fernández Universidad de Sonora [email protected]
Jesús Villalpando Quiñónez Boulder University [email protected]
According to Dooley (2010: 90), clause chaining is a term referring to “long sequences of foreground clauses” or “sequential event clauses”. The author also considers that the most important property of such sequences is the presence of switch-reference markers. On the other hand, Overall (2014) considers that clause chaining can be defined as a “multi-clausal construction consisting of one or
39 more dependent clauses associated with a single finite clause.” (p. 314). In line with the aforementioned definitions, this paper provides an analysis of clause chaining in Tarahumara, a Uto-Aztecan language from northwestern Mexico. The analysis that we develop assumes a functional corpus- oriented perspective and argues that the phenomenon of clause chaining in Tarahumara makes use of other operators rather than switch-reference markers (lacking in Tarahumara). The organization of sequences of clauses in Tarahumara is achieved by means of simple and complex conjunctions and subordinators, as it is shown in (1) and (2), as well as by a set of different nominalizing suffixes:
(1)Simple conjunctions a’rí ko échi kochí wirísi-ri, CONJ EMPH DEM dog stand_up-PFV a’rí chóta-ri échi músa neká-mia. CONJ start-PFV DEM cat bark-PURP EN: ‘And then the dog stood up and started to bark to the cat.’ ESP: ‘Y entonces el perro se levantó y empezó a ladrar al gato.’
(2)Complex conjunctions mapu-a’rí chóta-ri bacháwara ukí wichí-ya SUB-SIM start-PFV first rain fall-NMLZ Pedro simí-re iwé-chi Pedro go-PFV field-LOC a’rí karéwi-ri échi iwé kíti ichi-méa sunú. CONJ weed_out-PFV DEM field PURP sow-IRR corn EN: ‘When the first rains started to fall down, Pedro went to the field / and weed it out in order to sow corn.’ ESP: ‘Cuando comenzaron a caer las primeras lluvias, Pedro fue al campo y desyerbó la milpa para sembrar maíz.’
40 The nominalizing suffixes observed in clause chaining are: -a, - ka and -cho. These suffixes seem to belong to different functional domain: -a is relevant in complement clauses, -ka in some adjectival constructions, and -cho in adverbial clauses. Clause chaining in Tarahumara demonstrates that the boundaries between clauses at discourse level are not always clear and that clause types may overlap.
Referencias
Dooley, Robert A. 2010. Exploring Clause Chaining. SIL Electronic Working Papers in Linguistics. http://www.sil.org/ system/files/reapdata/13/04/21/130421360311569 160306235201738876375986/silewp2010_001.pdf [September 30, 2014] Overall, Simon E. 2014. “Clause chaining, switch reference and nominalisations in Aguaruna (Jivaroan).” In Information Structure and Reference Tracking in Complex Sentences. Rik van Gijn, Jeremy Hammond, Dejan Matic, Saskia van Putten and Ana Vilacy Gaucio, 309-340. Amsterdam: John Benjamins.
41 Oraciones adverbiales temporales y correferencia de sujetos Rebeca Gerardo Tavira Universidad Nacional Autónoma de México [email protected]
Las oraciones adverbiales temporales funcionan como escenario temporal o punto de referencia para el estado de cosas expresado mediante el verbo de la oración principal (Heinämäki 1974; Cristofaro 2003; García 2000). Uno de los factores que permite evaluar el grado de integración de las relaciones complejas es la correferencia de los participantes de ambos estados de cosas. En su trabajo tipológico sobre subordinación, Cristofaro (2003: 166) postula que rasgos como la referencia temporal, el valor aspectual y el valor modal deben estar predeterminados si se usan los conectores del tipo after, before y when, cuando éstos estable- cen relaciones de anterioridad, posterioridad y simultaneidad, res- pectivamente. En cambio, las oraciones temporales parecen no predeterminar la identidad de los participantes. Un estudio a partir de corpus del español permite observar algunas tendencias interesantes con respecto a la realización argumental de los participantes tipo sujeto, tanto en términos de correferencialidad como de codificación léxica. En trabajos pre- vios (García 2000) se ha dicho que, por ejemplo, antes y des- pués pueden introducir una oración de infinitivo con sujeto explicito, siempre y cuando este sujeto no sea correferencial con el sujeto de la oración principal, ej. Te fuiste antes de llegar yo. Además, si el sujeto es explícito, entonces no puede ser correferente con el sujeto principal:
(1) a. *Antes de entrar Juani en la habitación, __i se había dado cuenta de todo.
b. *Después de entrar Juani en la habitación, __i se dio cuenta de todo.
42 No obstante, los datos en un corpus amplio que incluye lengua oral y escrita (2000 oraciones, aproximadamente), se han podido comprobar las dos situaciones. En efecto, hay oraciones tempo- rales (en infinitivo y finitas) con sujeto explícito y distinto del sujeto de la principal (2), pero también hay casos en los que el sujeto explícito está en correferencia con el sujeto de la principal:
(2) a. Y antes de que Zoraidai pudiera hacerle ninguna otra pre-
gunta Matildej escapó (BC)
b. Ya después de que mi abuelitai falleció así como que __j dejé de ir a la iglesia (CM)
(3) a. Cástuloi, antes de __i avisar a la señora, __i se fue a la cocina seguido de Estefanía (RP)
b. Después de __i cenar, mi madrei, que __i está muy cansa-
da, __i fue a acostarse. (BC)
c. Don Roquei, el sacristán, después de __i bajar a los santos
__i se alejaba respetuoso (RP)
d. El visitantei antes de entrar al recinto __i debe firmar una
carta responsiva en la que __i manifiesta haber “leído y entendido el texto de SIN” (La Jornada)
Entre los factores que pudieran estar afectando la correferencia se han señalado el carácter factual o contrafactual de las oracio- nes, el orden en que aparecen las mismas, el tipo de relación (i.e., simultaneidad o secuencialidad) y el tipo de codificación del suje- to en ambas oraciones. El objetivo particular de este trabajo es explorar las posibles motivaciones que permiten (o no) la ocu- rrencia de sujetos explícitos, en oraciones temporales en infinitivo y finitas, establecer los patrones de correferencia (i.e. relaciones anafóricas), así como los contextos que presentan mayor ambi- güedad referencial, (i.e. referencia libre).
43 La alternancia de a y para en construcciones complejas con verbos de movimiento: ¿oraciones finales o de propósito? Paola Gutiérrez y Valeria Benítez Maestría en Lingüística Hispánica y Doctorado en Lingüística-UNAM [email protected] [email protected]
El criterio formal que ha predominado en la tradición hispánica para definir y clasificar construcciones subordinadas adverbiales es el nexo. Específicamente, las oraciones finales en español se identifican con las formas para (que) y a (que), aunque existen otros conectores menos prototípicos (Herrera 2002); estas cons- trucciones se definen como aquellas que “expresan el fin o la in- tensión con que se produce la acción del verbo principal” (Galán 1992), por ejemplo en hice un postre [para sorprenderte], salí al parque [a buscarte]. Se ha observado que a y para pueden, en muchos casos, alternar sin un cambio aparente de significado como sucede en corrí [a/para preguntarle el teléfono] (Gaviño 2009), sin em- bargo, esta posibilidad de permutación ocurre especialmente con verbos de movimiento y con algunos otros predicados cuya clase en común no está bien definida (ej. me ofrecí [?para/a ayudar- lo], me autorizó [?para/a venir]), aunque no siempre es posi- ble (ej. corrí [para/*a sentirme mejor]). De hecho, cuando las oraciones finales no codifican movimiento en la unidad principal, sino otro tipo de predicados, suelen aparecer con para(que), tal como se observa en ahorre mucho [para/*a viajar], me desve- lé [para/*a terminar]. En general, en la literatura poco se ha dicho sobre las construcciones complejas finales en corpus, y aún menos sobre las restricciones de orden y alternancia de nexo que estas estructuras pueden presentar (ej. Juan viene [a comer] [para/*a complacerte]).
44 En los trabajos tipológicos, por otro lado, las llamadas cláusu- las de propósito remiten a una definición semántico-conceptual (Cristofaro 2003; Schmidtke-Bode 2009), en la cual se subraya que un agente lleva a cabo una acción intencionalmente; además existe un vínculo entre dos eventos, de manera tal que el evento de la unidad principal se realiza con el objetivo, meta, intención de que el evento descrito en la unidad dependiente ocurra (Cristofaro 2003: 157). En apariencia, los rasgos de propósito coinciden con aquellos de las oraciones finales, sin embargo, una visión tipológica no consideraría algunas construcciones que se han tratado como finales en español, por ejemplo tengo la libertad [para esco- ger], vine a la biblioteca [para encontrarla cerrada], donde no hay un agente que actúe con volición o bien, el segundo evento no es deseado. Las construcciones con verbos de movimiento, en cambio, empatan mejor con la idea de propósito que se describe en muchas otras lenguas (Schmidtke-Bode 2009). Los objetivos de esta presentación son, primero, hacer una caracterización de las construcciones complejas que codifican mo- vimiento y de sus restricciones en la alternancia de los nexos a (que) y para (que); segundo, retomando los trabajos tipológicos, subrayar la pertinencia de distinguir en español entre propósito y finalidad, ya que el nexo no es un criterio suficiente para delimitar oraciones, pues sólo algunas de las llamadas finales responden a rasgos semánticos particulares (i.e. agente, volición, posteriori- dad, etc.). El estudio se basa principalmente en datos de lengua oral obtenidos de entrevistas y en pruebas de agramaticalidad.
Referencias
Cristofaro, S. 2003. Subordination. New York: Oxford University Press. Galán R., C. 1992. Las oraciones finales en español. Estudio sin- crónico. Cáseres: Anuario de Estudios Filológicos, Anejo No. 9.
45 Gaviño R., V. 2009. La finalidad como función lingüística. Oviedo: Septem Ediciones. Herrera, M.E. 2002. Nexos adverbiales en las hablas cultas y popular de la Ciudad de México. México: UNAM. Schmidtke-Bode, K. 2009. A typology of purpose clauses. Amsterdam: John Benjamins.
Comportamiento del verbo sentir en oraciones complejas: un estudio a partir de corpus Irasema Cruz Domínguez Universidad Nacional Autónoma de México [email protected]
Generalmente, se asume que los verbos de percepción codifican, de manera preferencial y primaria, la noción de percepción física; como tal, se espera que estos verbos aparezcan dentro de una estructura transitiva con dos participantes obligatorios: sujeto y objeto (Cano Aguilar 1987). En un estudio previo a partir de cor- pus oral, encontramos que las cláusulas simples con el verbo sen- tir muestran varias particularidades tanto a nivel sintáctico, como semántico. A partir de oraciones como las de (1) y (2), se encon- tró que las cláusulas simples con sentir aparecen preferentemente como oraciones intransitivas, con un solo participante: el sujeto experimentante. Se mostró también que sentir suele acompañarse de complementos predicativos para completar su significación, ej. sentirse alegre, sentirse abogados. En términos semánticos, es- tas estructuras suelen adscribirse al dominio emotivo (1) y, en menor medida, codificar percepción física o cognitiva (2). me siento una persona muy alegre/ muy cotorra/ o sea// a veces digo/ estupideces/ estupideces/ tontería y media y/ y surge la risa/ ¿no?/ [CSCM: 6, 176]
46 los compañeros que son jovencitos/ son los más egoístas/// son los que/ todavía ni siquiera terminan la carrera y ya se sienten abo- gados/// en serio/ [CSCM: 9, 222] Esta segunda fase de investigación tiene dos objetivos: (i) exa- minar si las características encontradas a nivel de la oración simple están presentes (o no) en construcciones complejas, y (ii) compa- rar el comportamiento de las cláusulas complejas con sentir como verbo matriz en un corpus amplio de lengua oral y lengua escrita.2 En términos sintácticos, sentir aparece en oraciones completivas introducidas con que (1) y, en menor medida, atribuciones, como el par en (2), usualmente denominadas cláusulas mínimas (Demonte y Masullo 1999). Además, también encontramos cláusulas completivas introducidas por nexos adverbiales, los cuales pro- porcionan información adicional al segundo evento (3).
(1)[…] siento [que no puede estar tomando pastillas de emergen- cia y pastillas de emergencia] le dije “pues no / porque después se va a hacer estéril [CSCM: 95, 306]
(2)a. no pude dejar de sentir [que la humanidad es como el ra- tón Miguelito cuando toma el libro de un mago y comienza a jugar con su inmenso poder.][CREA: Caminitos de plata, 2001] b. no pude dejar de sentir a la humanidad como el ratón Miguelito […]
(3) La voz de Agustín Lara inundó la recámara: Sol de mi vida Luz de mis ojos siente [cómo mis manos acarician tu tersa piel mis pobres manos.] [CREA: Tan veloz como el deseo, 2001]
2 Corpus Sociolingüístico de la Ciudad de México (CSCM) (Martín Butragueño y Lastra) y Corpus de Referencia del Español Actual (CREA).
47 En términos semánticos, el análisis de los datos sugiere que, tanto en lengua oral como escrita, sentir como verbo matriz se comporta como un predicado cognitivo de tal forma que codifica valores de actitud propositiva y/o epistémicos (6). Llama la aten- ción que, solo en el corpus textual, se documentan complementos proposicionales (certeza), como la expresión de lamento en (7).
(4) me siento orgulloso de lo que Dios me dio/ porque no es mi capacidad/ sino que yo siento [que Dios a cada uno nos pone un talento ¿no?] [CSCM: 75, 707]
(5) Explica que desde su juventud añoraba y sentía profundamente [que sus amigos y los jóvenes mexicanos no tuvieran un museo] [para inspirarse]. [CREA: Revista Digital Universitaria, v. 4, nº 1, 03/2003]
Entre pedir y el deseo. Gramaticalización del desiderativo en maayat’aan Fidencio Briceño Chel Centro INAH; Yucatán
En este trabajo presentaré datos del maya peninsular actual sobre la caracterización de predicados desiderativos en esta lengua mayance, por un lado evidenciaré la estructura sintáctica que mues- tra su proceso de gramaticalización y por otro caracterizaré su funcionamiento, las condiciones de uso y los cambios de significa- do originados por este proceso. Los datos aquí presentados nos ayudarán a mostrar que uno de esos predicados (k’áat) puede tener un sentido que va desde ‘querer’, ‘desear’, ‘anhelar’ hasta ‘esperar’, tal como ha sido especificado por Noonan en su clasificación de verbos (2007: 145);
48 sin embargo en yucateco observamos que el punto de arranque viene de un predicado con el sentido de ‘pedir’-‘querer’, que dependiendo de sus avances de gramaticalización desarrolla el sentido de ‘desear’, lo cual nos lleva a buscar la división semánticamente en tres clases planteado por Noonan (op. cit.) de acuerdo a su función: (a) ‘esperar’, (b) ‘desear’ y (c) ‘querer’. Por otro lado, mostraré que otro de los predicados con senti- do desiderativo (Taak) muy probablemente parte de construccio- nes donde está presente el deseo de una necesidad fisiológica, pero que con el paso del tiempo y a través de su gramaticalización como desiderativo empieza a usarse con otros verbos que no im- plican una necesidad fisiológica inmediata. Con esto pretendo mostrar que a pesar de haber cierta coin- cidencia semántica entre oraciones desiderativas formuladas con k’áat o con taak, la estructura sintáctica evidencia distintos mo- mentos de gramaticalización y por otro también veremos que el sentido de usar uno u otro muestra la posibilidad de referir una necesidad inmediata o un deseo a futuro.
Referencias
Berbeira Gardón, José Luis. 1998. Dimensiones pragmáticas de la gramatización, Valencia, Universidad de Valencia. Bybee, J. L. y W. Pagliuca. 1985. “Cross-linguistic comparison and the development of grammatical meaning”, Historical Semantics. Historical Word Formation, J. Fisiak, ed., Amsterdam Mouton, 59-83. Noonan, M. 2007 [1985]. “Complementation”. En: T. Shopen (Ed.). Language Typology and Syntactic Description. Volume. II: Complex Constructions. Cambridge: Cambridge University Press. 52-150. Verhoeven, Elisabeth. 2007. Experiential constructions in Yucatec Maya, Amsterdam/ Philadelphia: John Benjamins.
49 The syncretism between antipassive and causative in Mocovi Cristian Juárez Universidad de Sonora [email protected]
Albert Álvarez González Universidad de Sonora [email protected]
In terms of valency changing operations, antipassive and causative constructions are considered as opposite since the antipassivization process corresponds to a valency decreasing operation whereas the causativization process implies a valency increasing (Dixon and Aikhenvald 2000; Givón 2001; among many others). Therefore, the same marker is not expected to be involved simultaneously in both processes. This presentation will show that the syncretism antipassive/ causative is recognized by the use of the verbal marker -(a)an in Mocovi, a Guaycuruan language spoken mainly between Chaco and Santa Fe provinces in Argentina. Although such syncretism seems to be unusual cross-linguistically, it is also attested in some West Mande languages, such as Creissels (2012) reports. Firstly, we will describe how the -(a)an suffix is used to express antipassive clauses and to also mark causative clauses in Mocovi. Secondly, we will explain the syncretism antipassive/ causative, considering the functional similarities between both types of valency mechanisms. Finally, we will adopt a diachronic perspective in order to determine the source of this verbal marker and to describe the grammaticalization pathway that has originated the syncretism currently observed.
50 References
Creissels, D. 2012. The origin of antipassive markers in West Mande languages. Paper presented at the 45th Annual Meeting of Societas Linguistica Europea, Stockolm, 29 August-1 September. Available in http://deniscreissels.fr Dixon, R.M.W. and Alexandra Aikhenvald (Eds.). 2000. Changing Valency. Case studies in Transitivity. Cambridge: Cambridge University Press. Givón, T. 2001. Syntax I. An introduction. Amsterdam: John Benjamins.
51 CONFERENCIA
53 Nominalization, de-subordination and re-finitization* T. Givón Linguistic Department University of Oregon and White Cloud Ranch Ignacio, Colorado [email protected]
1. Background1
1.1. The Uto-Aztecan continuum
Across the Uto-Aztecan family, one finds a sharp distinction between the extreme nominalizing north and the resolutely finite south. The two northern-most sub-families, Numic and Takic, nominalize every subordinate clause in sight. The same seems to
* In the preceding two chapters we noted that in some languages all subordinate clauses are nominalized, and thus exhibit less-finite features. In this chapter it is suggested that the nominalization of subordinate clauses, however natural in itself, leads to certain synchronic anomalies in terms of language processing strategies. Speakers devote one processing strategy to the most frequent clause-type in discourse, main clauses, but must change their processing strategy, often radically, to process subordinate clauses. It is suggested here that the process of re-finitization of nominalized subordinate clauses, which seems to occur eventually in many nominalizing languages, is prima facie evidence to this synchronic processing anomaly. What this chapter attempts to do is, first, discuss another diachronic process, the de-subordination of subordinate clause, and how it might intersect with re-finitization. It then probes the diachronic mechanisms by which re-finitization of subordinate clauses may take place. 1 An early version of this paper was presented at the CNRS Symposium on Finiteness and Nominalization, Paris, Sept. 2011. I am grateful to the
55 be true of Yaqui and Huichol. But further south one finds uniformly finite subordinate clauses in Tepiman (Tepehuan, Pima Bajo) all the way to Nahuatl. The transition zone between the two extre- mes, Guarijío, Trahumara and perhaps Cora, is surprisingly thin. There are good reasons for suggesting that the Uto-Aztecan north is both culturally (hunting-gathering) and linguistically (OV syntax) more conservative, and that the family’s south, due to either natural drift or contact with the Meso-American substratum is more innovative. So the question I would like to pose here is this: How does the drift from nominalized to finite subordinate clauses take place? Especially natural drift that is not induced by contact. The strategy I will pursue here is two-fold:
• Try to understand the internal logic of nominalization and related processes. • Try to find evidence, in the middle zone of the family and elsewhere, for the dynamics of change.
1.2. Finiteness and nominalization
At first glance it is tempting to assume a logical identity, i.e. bi- conditional, between nominalization and non-finiteness, as in:
(1) NOM NON-FIN
organizers, Claudine Chamoreau and Zarina Estrada-Fernández, as well as the participants, for the most stimulating discussion and enjoyable ambiance. An expanded version of the paper was presented at the Semi- nario de Complejidad Sintáctica, Universidad de Sonora, Hermosillo, November 2014. Special thanks are due to Scott DeLancey for sharing and discussing the data on Tibeto-Burman nominalizations.
56 Excess caution may suggest a non-committal partial overlap, as in (Bisang 2011):
(2) NOM NON-FIN
What I suspect the data will support better is a proper inclusion relation, i.e. one-way conditional, as in (3) below, whereby nominalized clauses are always non-finite, but some less-finite structures are not necessarily the product of nominalization:
(3) NOM NON-FIN
While finiteness is often discussed as a property of verbs, a more comprehensive view would recognize it as a property of clauses. If one takes the main-declarative-affirmative-active clauses to be the prototype of the finite verbal clause, then its syntactic properties may be viewed as the benchmark from which less-finite clauses, whether nominalized or not, may diverge. Such divergence moves the nominalized clause towards another prototype, that of the nominal phrase (NP). And one may then view nominalization as a diachronic derivational process via which the prototype finite verbal clause is converted, as close as it can go, to the prototype of nominal phrase. That is:
(4) Nominalization as a diachronic process: Nominalization is the process via which a finite verbal clause-either in its entirety or just a subject-less verb phrase- is converted into a noun phrase.
Nominalized or non-finite clauses can be then described in terms of the structural adjustments that apply to the clause upon its conversion from the finite/verbal to the non-finite/nominal prototype (Hopper and Thompson 1984). The major components of such adjustment are:
57 (5) Structural adjustments from verbal to nominal prototype: finite verbal clause nominalized (non-finite) clause ======a. lexical category: verb > head noun b. nominal marking: ——— > nominal marking on verb c. verbal modalities: T-A-M > loss or reduction d. pronouns: pronominal AGR > loss or reduction e. case marking: subject/object > genitive f. restriction: ——— > added determiners g. modification: adverbs > adjectives h. clausal marking: ——— > case-marking on clause
The inventory of syntactic features associated with clause nominalization as given in (5) is the expanded inventory of all possible features. As elsewhere in typology, individual languages may display the entire set or just a sub-set of features (5a-h), giving rise to a more fine-grained typology. Presumably, by comparing the cross-language distribution of these features one may arrive at an implicational-hierarchic scale that would tell us which of these features are more necessary and universal, and which are more optional and ancillary. And as often happens in synchronic implicational hierarchies, the scale may turn out to reflect the diachony. But why should one want to nominalize a clause to begin with? The answer to this may be given at several levels. At the most concrete level, one may rephrase the question as:
• In what syntactic or communicative context(s) does one most commonly find nominalized clauses?
The most obvious answer to this question is that a verbal clause is nominalized most commonly in contexts where it occupies a
58 prototypical nominal position, that of subject, direct object, indirect object or nominal predicate inside another clause. In other words, nominalization tends to march in tandem with clausal embedding and syntactic complexity. As an illustration of the structural pattern that emerges out of (5) above, contrast the finite clause (6a) below with its non-finite nominalized counterpart (6b):
(6)a. Finite verbal clause: She knew mathematics extensively b. Nominalized NP: Her extensive knowledge of mathematics
The same tradition that views finiteness as a property of verbs also treats it as a discrete either-or phenomenon. But since the finite verbal prototype, or its nominal converse, is defined by multiple features (5a-h), finiteness is in principle a matter of degree. Consider, for example, the graduated finiteness scale in English, as in (7) below. At the very bottom, one finds the prototype finite clause (7h). At the very top, one finds the radically nominalized non-finite clauses (7a). But most clause-types on scale (7) fall in- between, exhibiting intermediate degrees of finiteness and nominalization.
(7) least finite (nominalized) ======a. [Her good knowledge of math] surely helped b. [Her knowing math well] surely helped c. [For her to know math so well] surely helped d. She wanted [to know math well] e. [Knowing math well], she then... f. [Having known math well since childhood], she... g. He assumed [(that)she knew math well] h. She knew math well ======most finite
59 2. Typological variation in the distribution of finite vs. nominalized structures
2.1. ‘Permissive’ languages
English is a rather convenient language to open the discussion of nominalization with, in that it exhibits a whole range of structures, from the fully finite (7h) to fully nominalized (7a), all presumably used in their proper syntactic-communicative contexts. What English also illustrates is that in a single language, and in the very same grammatical contexts, one can have the option of using either a more finite or a more nominalized structure. Thus consider:
(8)context finite nominalized (less finite) ======a) ADV-clause: After she returned After her return home home b) V-complement: He knew she loved He knew of her love flowers of flowers c) V-complement: I hope I can do it I hope to do it tomorrow tomorrow d) V-complement: I expect that he I expect him to do it would do it e) Clausal subject: That she loved Her love of flowers flowers surprised surprised him him f) REL-clause: She was looking She was looking for for someone someone to love she could love
60 What examples (8) illustrate is that a language may have con- siderable leeway in using variant levels of finiteness in roughly-the- same syntactic contexts, presumably yielding more subtle semantic or pragmatic effects. Often, such variation points either to emerging diachronic expansion of syntactic patterns, or to surviving vestiges of older patterns that are being modified and reshaped. Languages like English, whether due to an innovative or reductive diachronic trajectory, are interesting because they illustrate the mid-range of the finiteness typological continuum, exhibiting a large variety of finite and less-finite structures. Such variability within the same language is found in families such as Indo-European, Semitic, Bantu and others2. But there are also languages –and language families– that occupy either extreme position on the finiteness scale. We will survey one such extreme directly below, having surveyed an example of the other extreme in ch. 25, above (see also vol. II, ch. 15).
2.2. Extreme nominalizing (embedding) languages
The extreme-nominalizing language type is found in a number of language families or sub-families-Tibeto-Burman (Watters 1998, 2008; Genetti et al. 2008; Hyslop 2011; DeLancey 2011), Turkic (Lewis 1967), Carib (Gildea 1998), Quechuan (Weber 1996), Gorokan (Thurman 1978), and No. Uto-Aztecan (Hill 2005; Givón 2011). We will illustrate this type with Ute (Numic, Uto-
2 In two of these families, Semitic and Indo-European, there is internal evidence to suggest that the typology of finiteness has probably shifted— perhaps even back and forth—over historical time. In Bantu likewise, an older extreme-finite (serial-verb) syntax was most likely re-shaped into the current English-like mix of finite and less-finite subordinate clauses (Givón 1975; see vol. I, ch. 7).
61 Aztecan). The three most conspicuous structural features of clause nominalization in Ute are:
• nominalizer suffix on the verb (5b) • genitive case-marking of the subject (5e) • object case-marking of the entire clause (5h)
Compare first the finite verbal clause (9a) below with its nominalized counterpart (9b), serving as the subject of another clause, and (9c), serving as another clause’s object:
(9) a.Finite main clause: ta’wachi ‘u yoghovchi pakha-ukh-kwa man/S the/S coyote/O kill-go-ANT ‘the man killed the coyote’ b. Nominalized clausal subject: [‘uru ta’wachi ‘uway yoghovchi [that/O man/G the/G coyote/O pakha-ukh-kwa-na]t-’ay kill-go-ANT-NOM/S] good-IMM ‘it is good that the man killed the coyote’ (hist.: ‘the man’s killing (of) the coyote is good’) c. Nominalized clausal object: puchuchugwa-qha [‘uru ta’wachi ‘uway know-ANT [that/O man/G the/G yoghovchi pakha-ukh-kwa-na-y] coyote/O kill-go-ANT-NOM-O] ‘(she) knew that the man killed the coyote’ (hist.: ‘(s/he) knew the man’s killing (of) the coyote’)
The nominalized clauses in (9b, c) show at least one conspicuous element of finite structure -tense-aspect-modality. Indeed, many historically-nominalized clauses in Ute exhibit this finite feature, an issue we will return to further below.
62 The very same nominalized structure is found in Ute object REL-clauses:
(10)yoghovchi ‘u [ta’wachi ‘uwáy coyote/S the/S [man/G the/G pakha-khwa-kwa-na]-y pnikya-qhay-’u kill-go-ANT-NOM]-O see-ANT-3s ‘(I) saw the coyote that the man killed’ (hist.: ‘(I) saw the coyote of the man’s killing’)
In equi-subject complements of modal-aspectual and manipulation verbs, the verb is marked with either of two other nominalizing suffixes, and the only T-A-M marking allowed is the irrealis modality:
(11) a. Finite main clause: na’acichi‘u tkuavi ‘uru tka-qha girl/S the/S meat/O the/O eat-ANT ‘The girl ate the meat’ b. SS complement of modal-aspectual verb: na’acichi‘utkuavi ‘uru tka-vaa-chi girl/S the/S meat/O the/O eat-IRR-NOM/SS ‘ásti’i-kya want-ANT ‘the girl wanted to eat the meat ‘ (hist.: ‘the girl wanted (the) eating of the meat’) c. DS complement of manipulation verb: mamachi ‘u na’acichi ‘uway tkuavi ‘uru woman/S the/S girl/O the/O meat-O the/O tka-vaa-ku máy-kya eat-IRR-NOM/DS tell-ANT ‘the woman told the girl to eat the meat ‘ (hist.: ‘the woman told/ordered the girl the eating (of) the meat’)
63 Adverbial clauses are also nominalized in Ute, using the same DS suffix -ku used in complements of manipulation verbs (11c). Some ADV-clauses carry no T-A-M marking, which is predictable from the associated main clause. Thus consider:
(12) a. Realis ‘when’-clause: ta’wachi ‘uway kani-naaghayga-khu-’uru, man/G the/G house-in enter-NOM-that mamachi ‘u págha-kwa-qha woman/S the/S go-go-ANT ‘when the man entered the house, the woman took off’ (hist.: ‘(upon) the man’s entering the house, the woman took off) b. Irrealis ‘if’/’when’ clause: ta’wachi ‘uway kani-naaghayga-khw-’uru, man/G the/G house-in enter-NOM-that mamachi ‘u págha-kwa-vaa-ni woman/S the/S go-go-IRR-FUT ‘if/when the man enters the house, the woman will take off’ (hist.: ‘(upon) the man’s entering the house, the woman will take off’)
The anterior-perfect aspect can also appear in ADV-clauses, with the predictable meaning of subsequence (‘after’). Thus, com- pare (13) below to (12a) above:
(13) ta’wach-i ‘uway kani-naaghayga-qhay-khu-’uru, man-G the/G house-in enter-ANT-NOM-that mamach ‘u págha-kwa-qha woman/S the/S go-go-ANT ‘After the man entered the house, the woman took off’ (Lit.: ‘(following) the man’s having entered the house, the woman took off)
64 And the irrealis marker can be added to the anterior-perfect, yielding a counter-fact conditional sense, as in:
(14) ta’wachi ‘uway kani-naaghayga-qha-vaa-ku,... man/G the/G house-in enter-ANT-IRR-NOM ‘if the man had entered the house (tho he didn’t)...’
As noted earlier (ch. 26), subject REL-clauses in Ute display their own pattern of nominalization, with the subject/agent nominalizer -t and no restrictions on finite T-A-M marking. Thus consider:
(15) a. Finite main clause: ‘áapachi‘utkuavi ‘uru tka-qha boy/S the/S meat/O the/O eat-ANT ‘the boy ate the meat’ b. Subject REL-clause: ‘áapachi‘utkuavi tka-qa-t pnikya-qha boy/S the/S meat/O eat-ANT-NOM/O see-AND ‘(I) saw the boy who ate the meat’ (hist.: ‘(I) saw the meat-eating boy’) c. Nominalized predicate: ‘áapach ‘u tkua-tka-mi-t ‘ura-’ay boy/S the/S meat-eat-HAB-NOM be-IMM ‘the boy is a meat-eater’
The same subject- nominalizer -t is used widely in other subject/ agent nominalizations, as in:
(16)a. Possessor: puwa-gha-t ‘medicine-man’ power-have-NOM b. Negative possessor: ka-p’i-’a-t ‘blind person’ NEG-eye-have/NEG-NOM’
65 c. Subject-of -passive (old form): p’-kwa-t ‘book’ (hist. ‘what was written’) write-PASS-NOM d. Adjective: ‘aka-gha-r ‘red’ (inan.; hist: ‘that has redness’) red-have-NOM
In one hard-to-characterize subordinate construction, the main- clause verb is nominalized with the subject nominalizer -t, and is then subordinated, perhaps as subject, to the finite main verb ‘be’. Some typical text-derived examples of this are3:
(17) a....nachuwa-pi ‘apagha-pi ‘úru, ‘úru ‘ava’na fancy-NOM/O talk-NOM/S that/S that/S much ‘ura-t ‘ura-yi-s ‘iya-na... be-NOM be-IMM-C here-LOC ‘...talking fancy talk, there’s lot of that here...’ b....míiya-ni-naagha ‘uni-’ni-kya-t-m ‘ura-yi-s... far-LOC-maybe do-INT-PL-NOM-PL be-IMM-C ‘...(or) maybe those who live far...’ c. ...p-pa kar-khwa-ta radio ‘uru turn-ony-ta, road-DIR sit-go-NOM radio that/O turn-on-NOM pa’a-’ura-t-s ‘ura-mi... complete-be-NOM-C be-HAB ‘...driving on the road (through their reservation) and turning the radio on, it is completely that (speaking Navajo on the radio)...’ d....’ava’na-aqh ‘uni-kya-t-m ‘ura-’ay... many-it do-PL-NOM-PL be-IMM ‘...they do a lot of it...’
3 Talk by the late Harvey Natchez (Uncompaghre, No. Ute) to the Tri- Ute Language Conference, March 30, 1979.
66 e. ...puchuchugwa-vaa-ni-’uru, ‘át-mani-kya-t-m know-IRR-FUT-that well-do.like-PL-NOM-PL ‘ura-yi-s... be-IMM-C ‘...they will (be able to) learn and do well like this...’
Lastly, as noted earlier (vol. I, ch. 16), the passive in Ute is also a diachronic product of nominalization, in this case of a subjectless VP. Thus compare:
(18)a. Active-transitive: ‘áapachi‘u tkuavi ‘uru tka-qha boy/S the/S meat/O the/O eat-ANT ‘the boy ate the meat’ b. Passive: tkuavi ‘uru tka-ta-qha meat/O the/O eat-PASS-ANT ‘the meat was eaten’, ‘someone at the meat’ c. Action nominal: tka-ta t’a-t ‘ura-’ay eat-NOM good-NOM be-IMM ‘eating is good’ d. VP nominalization: tkuavi ‘uru tka-ta t’a-t ‘ura-qha meat/O the/O eat-NOM good-NOM be-ANT ‘eating the meat was good’
One may suggest that in extreme nominalizing languages such as Ute, the feature subordinate clause is grammaticalized to the max via nominalization. Such a strategy is akin to treating subordinate clauses, by analogy, as subject or object nominals.
67 2.3. Extreme finite languages
At the other extreme of the typological continuum one finds languages in which all clause-types are finite, including, in some languages, even lexical nominalizations. Several Amerindian families display this phenomenon, e.g. Iroquois, (Mithun 1991), Algonquian, Siouan-Cadoan, So. Arawak, and Athabaskan. The same may be seen in many serial-verb languages in Southeast Asia (Thai, Burmese, Mon-Khmer, Yao-Miao; Lahu, Matisoff 1972) and Africa (Akan, Osam 1994; Senufu, Carlson 1994; and many other Niger-Congo languages; see vol. I, ch. 7). An extensive illustration of this extreme type with data from Tolowa Athabaskan may be found in vol. I, ch. 15, as well as ch. 25 above and ch. 29, below.
3. The diachronic logic of clause nominalization
As noted earlier above (see also chs 25, 26), the underlying logic of clausal nominalization is most likely analogical:
(19)If a clause occupies a prototypical nominal position, it may be treated syntactically as a noun phrase.
The logic of (19) seems fairly straight-forward in V- complements, whose main verbs are originally transitive verbs that take nominal objects. Extending their semantic scope to modal- aspectual, manipulation or perception-cognition-utterance senses is ubiquitous (Dixon 1991). Thus consider:
68 (20)old sense with new sense with nominal object clausal complement ======Modal-aspectual verbs: She wanted an apple > She wanted to eat an apple She stopped the car > She stopped driving the car She avoided the bridge > She avoided crossing the bridge Manipulation verbs: He told her a storey > He told her to leave He stopped her > She stopped her from leaving He wanted her > He wanted to marry her P-C-U verbs: They saw him > They saw him working > They saw he was working He knew her > He knew she was there She told him a story > She told him that she was busy
The logic of nominalized REL-clauses is less obvious. While modifying REL-clauses are embedded in a noun phrase, their position as restrictive modifiers is more akin to that of adjectives or quantifiers, i.e. predicates. The logic of nominalizing REL- clauses becomes more transparent, however, as noted earlier (see ch. 26), the early source of modifying REL-clauses is often a headless REL-clause used first as a paratactic, non-restrictive REL-clauses. Thus compare:
(21)a. Finite main clause: ta’wachi ‘u yoghovchi ‘a-qa man/S the/S coyote/O trap-ANT ‘the man trapped the coyote’ b. Headless REL-clause: puchuchugwa-y [‘uway yoghovchi ‘a-qa-t] know-IMM-3s that/O coyote/O trap-ANT-NOM/O ‘(I) know the one who trapped the coyote’ (hist.: ‘I know that coyote trapper’)
69 c. Non-restrictive REL-clause: puchuchugwa-y ta’wachi ‘uway, [‘uway know-IMM man/O the/O that/O yoghovchi ‘a-qa-t] coyote/O trap-ANT-NOM/O ‘(I) knows the man, the one who trapped the coyote’ (hist.: ‘I know the man, that coyote trapper’) d. Restrictive subject REL-clause: puchuchugwa-y ta’wachi ‘uway yoghovchi know-IMM-3s man/O the/O coyote/O ‘a-qa-t trap-ANT-NOM/O ‘(I) know the man who trapped the coyote’
Likewise for object REL-clauses:
(22)a. Finite main clause: mamachi ‘u páanay ‘uru chíir’a-qa woman/S the/S bread/O the/O fry-ANT ‘The woman fried the bread’ b. Headless REL-clause: pnikya-qhay-ku [‘uru mamachi chíir’a-qha-na-y] see-ANT-it that/O woman/G fry-ANT-NOM-O ‘(I) saw what the woman fried’ (hist.: ‘I saw that of the-woman’s-frying’) c. Non-restrictive REL-clause: pnikya-qhay-ku páanay, [‘uru mamachi see-ANT-it bread/O that/O woman/G ‘uway chíir’a-qha-na-y] the/G fry-ANT-NOM-O ‘(I) saw the bread, which the woman fried’ (hist. ‘I saw the bread, that of the woman’s frying’)
70 d. Restrictive REL-clause: pnikya-qha páanay [‘uru mamachi see-ANT bread/O that/O woman/G ‘uway chíir’a-qha-na-y] the/G fry-ANT-NOM-O ‘(I) saw the bread that the woman fried’
4. De-subordination
I have earlier defined re-finitization as the acquisition of finite features by an erstwhile nominalized clause (Givón 1993). Mithun (2011) has suggested that the phenomenon called by Evans (2007) insubordination is also an instance of re-finitization. It seems to me that we have two distinct phenomena here that arise via different diachronic pathways, under different functional-adaptive pressures, and thus require distinct description. The synchronic phenomenon described by Evans (2007) as ‘insubordination’, and that earlier I called the ‘re-surfacing’ or ‘liberation’ of erstwhile subordinate clauses, may be better characterized in diachronic terms as de-subordination. This is a process whereby erstwhile subordinate clauses, often nominalized and/or less finite, become main clauses. A most succinct description of this process in Tibeto-Burman languages has been given by DeLancey (2011): “...TB languages frequently innovate new, marked clausal constructions with nominalized verbs and finite copulas...Frequently such constructions lose their marked status and become ordinary finite constructions...Many TB verbal systems transparently reflect this origin, for example modern Tibetan tense/aspect forms like - pa-yin, -pa-red, both consisting of the nominalizer -pa in construction with an equational copula...” (2011, p. 1) The grammaticalizaion of main verbs as T-A-M markers is a common mechanism in de-subordination, often resulting in the
71 introduction of other syntactic features into the de-subordinated main clause. Thus, for example, both in Tibeto-Burman and Cariban de-subordination has introduced ergative case-marking, in Cariban by re-analyzing erstwhile dative-marked agents (Gildea 1998), in Tibetan by re-analyzing erstwhile genitive-marked agents. In Japanese, the current -ga subject marker in finite main clauses is a re-analyzed genitive of earlier nominalized subordinate clauses (Akiba 1978; Shibatani 2007). And in No. Uto-Aztecan the de- subordination of nominalized VPs is responsible for the merger of object and genitive case-marking in main clauses (see vol. II, ch. 18). As an illustration of the latter mechanism, consider the grammaticalization of the remote-past marker -pga in Ute, derived from the combination of the inanimate/object noun-suffix -p, used in many types of nominalization, and the irregular–object incorporating–verb -ga ‘have’:
(23) a. Possessed noun: kani-gya-y ‘(s/he) has a house’ house-have-IMM b. Nominalized object:tka-p ‘eating’ eat-NOM c. Possessed nominalized verb:tka-p-ga-y ‘(s/he) has eating’ eat-NOM-have-IMM d. Re-analyzed as remote-past:tka-pga ‘(s/he) ate’ eat-REM
In the same vein, the current English present-progressive is the product of the de-subordination of a nominalized VP, preserving the grammaticalized copula ‘be’ and the nominalizer -ing as telltale traces.
72 While tense-aspect renovation is a common mechanism driving de-subordination, other mechanisms also exist. A common one is the re-interpretation of complement clauses of modality or manipulation verbs as independent main-clauses that perform deontic (subjunctive, hortative) speech-acts (Givón 1971b). In this process, the main verb is simply elided and its modal value is assumed by the erstwhile complement clause. Thus, in Ute:
(24) a. Complement of modality verb: tka-vaa-chi ‘ásti-’i eat-IRR-NOM want-IMM ‘(I) want to eat’ b. Subjunctive main clause: tka-vaa-chi-n eat-IRR-NOM-1s ‘I intend to/should/might eat’ c. Hortative main clause: tka-vaa-chi-rami eat-IRR-NOM-2d/INCL ‘let’s (you and I) eat’ d. Complement of manipulative verb: tka-qha-paa-ku máy-pga-am eat-PL-IRR-NOM tell-REM-3p ‘(I) told them to eat’ e. Hortative main clause: tka-qha-paa-ku-am eat-PL-IRR-NOM-3p ‘they should eat’, ‘let them eat’
A similar mechanism converts conditional ADV-clauses into manipulative (‘indirect’) speech-acts, eliding their main clauses, as in English:
73 (25) a. Subordinate ADV-clause: It would be nice if you could get me a spoon... b. Indirect speech-act: Now if you could get me a spoon?
Another mechanism, reminiscent of tense-aspect development, involves the grammaticalization of the main verb ‘be’, initially used as an emphatic cleft-like devise. When ‘be’ becomes grammaticalized, the construction loses its emphatic sense and become de-marked. Something like this must have happened in Kikuyu, where the old Bantu copula -ni is now the generalized marker of simple declarative clauses. The Ute emphatic construction cited in (17) above has the potential for developing in this direction. Thus, recall (17d) above, reproduced as4:
(26)...’ava’na-aqh ‘uni-kya-t-m ‘ura-’ay... many-it do-PL-NOM-PL be-IMM ‘...they do a lot of it...’
Lastly, a more complex mechanism of de-subordination may be seen in the recruitment of participial clauses into the grammar of clause chaining. Such clauses start their life as noun modifiers inside the NP. Next they become participial adverbial clauses, and eventually chain-medial same-subject (SS) clauses. The early stage of this development was noted in Latin (Haiman 1983). A more
4 Ute has grammaticalized the same ‘be’, -’ura or -’ara, as a topicalizing suffix on nouns, marking important discontinuous referents when coming back into the discourse, either at chain-initial or at chain-medial (switch- reference) position. This suffix is also part of clause-initial adverbial connectives (Givón 2011, ch. 18). The grammaticalization of ni ‘be’ in Kikuyu, first as an emphatic/cleft marker, then as a de-marked declarative- clause marker, is an instance of the same process.
74 advanced stage has been reported in Jiwarli (Austin 1992). Schematically, the process may be illustrated with English data as (see Givón 2001, ch. 18):
(27) a. Participial noun modifier: The running man took off b. Participial adverbial: The man, running, took off The man took off running Running, the man took off c. Chain-medial participial (OV pattern): Coming into the room, looking around and seeing nobody there, she relaxed. d. Chain-medial participial (VO pattern): She came into the room, looking around, seeing nobody, then relaxing.
Through all these diachronic pathways, the de-subordinated clause pulls into its new-found main-clause status whatever non- finite structural features it had as a subordinate clause. The subsequent addition of more finite features, what I would like to call re-finitization, is a separate process. Indeed, de- subordinations itself most commonly leads to the introduction of non-finite, nominalized features into main-clause syntax. So in a strict sense it does the very opposite of re-finitization.
5. Re-finitization
5.1. Preliminaries
The re-interpretation of de-subordinated clauses as the new stan- dard of finite main clauses, described by DeLancey (2011), is undoubtedly an important phenomenon. What I would like to focus
75 on here, however, is a different diachronic process–the refinitization of nominalized subordinate clauses. In attempting to account for the distribution of nominalized, non-finite morpho-syntax across the various types of subordinate clauses, one may consider invoking two alternative explanations to the seeming association between nominalization and subordinate clauses, one diachronic-syntactic, as in (19) above, the other communicative:
(28) a. Diachronic syntactic explanation: Nominalized syntax occurs in subordinate clauses when they occupy a prototypical nominal position inside main clauses. b. Communicative explanation: Non-finite syntax occurs in subordinate clauses that show maximal referential and tense-aspect-modal continuity (coherence) vis-a-vis their main clause.
These two explanations are not exclusive of each other, and the communicative (28b) may help explain, perhaps even in some way drive, the diachronic-syntactic (28a). For example, the nominalization of complements of modal-aspectual and manipulation verb does signal higher referential, temporal and thematic integration of main and complement clauses (see ch. 25, above). Likewise, nominalized or participial adverbial and chain-medial clauses exhibit strong referential and tense-aspect continuity vis-a-vis their main clauses (see examples (12) and (27) above). Still, for the moment it might be helpful to consider the two explanations separately. As suggested earlier above, clause nominalization is best understood as a diachronic process that, whatever its original motivation, winds up creating grammatical patterns that distinguish subordinate clauses from main clause—and from each other. Of the eight structural features that characterize nominalized clauses (5), one in particular—tense-aspect-modal marking (5c)—seems to vary greatly among the various types of subordinate clauses. Is
76 such variation predictable? And if so, by what logic? And does that logic operate during the original diachronic process of nominalization, or during a subsequent process of re-finitization? Table (29) below present a tentative ranking of the likelihood of T- A-M marking in the various types of subordinate clauses in English, a mid-range language. All such clauses would be nominalized in an extreme nominalizing language like Ute.
(29) Likelihood of finite T-A-M marking in subordinate clauses: least likely examples ======lexical nominalizations Knowledge is power —————————————————————— a. modality complements (equi-S) She wanted to leave b. manipulation complements (equi-O) She told him to leave c. purpose clauses (equi-S) She wet to school to study math d. participial clauses (equi-S) Running late, she took a cab (equi-O) She saw him running —————————————————––————— e. ADV clauses (SS) After he left, he called a taxi (DS) After he left, she called a taxi —————————————————————––— f. C-P-U verb complements, He knew she would rather leave g. REL-clauses The man you’ll meet was my teacher h. clausal subjects It’s a pity he didn’t show up ======most likely
77 The first thing to note is that the scalar continuum in (29) does not only rank the likelihood of finite T-A-M marking in a nominalizing language like Ute, but also the likelihood of using the nominalization strategy in a mid-range language like English. At the bottom of the scale (29e,f,g,h) are clause types that are not commonly nominalized in English. In Ute, those are precisely the nominalized clauses that can carry finite T-A-M marking. Next to notice is that the syntactic position of a subordinate clause within the main clause, our structural explanation (29a), is not the best predictor of whether it will be nominalized or non- finite. A much better predictor is the functional explanation (29b)— degree of continuity/coherence between main and subordinate clause, reducible to two major grammar-coded factors (Givón1983):
• referential continuity • tense aspect-modal continuity
The subordinate clauses at the bottom of the scale (29e-h) are those that do not require referential continuity, and can have T-A- M values independent of their main clause5. Indeed, much of the expressive power of these subordinate clauses rests in their ability to diverge from their main clause in these two core aspects of thematic continuity. In one clause-type, ADV-clauses (29e), the two possible predictors—continuity and syntactic position—make identical claims. Unlike V-complements and REL-clauses, ADV clauses do not occupy an obvious nominal position (S, O, IO, PRED)
5 Ute allows only two T-A-M markers in ADV-clauses, anterior-perfect -ka and the irrealis -vaa, but no tense marking (immediate, remote, habi- tual).
78 inside their main clause. Indeed, syntactically they are not part of the main clause at all, and are most typically packaged under a separate intonation contour. They also don’t require referential or T-A-M continuity with the main clause, except in the case of one sub-type—participial clauses (see further below). Notice, however, that while optional in both English and Ute, referential and T-A-M continuity between main and adverbial clauses often is the case, giving rise to less-finite ADV clauses in languages such as English. Thus contrast the finite ADV-clause with fully independent reference and T-A-M in (30a), below, with the non-finite ones with zero-marked reference and T-A-M in (30b):
(30) a. Finite ADV clause: After she came over to the house, he left. (> she came, he left) b. Non-finite ADV clause: After coming over to the house, she left. (> she came, she left) Having come over to the house, she left. (> she came, she left)
Since in the rest of the examples in (29) continuity suffices as a predictor, perhaps in (29e) too syntactic position is not the real predictor. Similar observations were made by Watters (2008) about more finite vs. less finite nominalized complement clauses in Kham (Central Himalayan, Tibeto-Burman), correlating the degree of finite subject/object agreement marking with the degree of referential continuity. Whatever factor predicts the ranking of subordinate clauses in (29), we still need to determine at what diachronic stage it operates:
• during the original stage of nominalization, where it may shield clause-types (29e-h) from loss of T-A-M marking; or • during a subsequent stage of re-finitization.
79 To answer this question fully would require finding and comparing languages in the early stages of nominalization, or in the early stages of re-finitization6.
5.2. Re-finitization and tense-aspect-modal marking
When I first played with the concept of re-finitization in (Givón 1993), I proposed that the appearance of T-A-M marking in Ute REL-clauses, ADV-clauses and verbal complements should be interpreted diachronically the following way:
(31) a. Stage I–nominalization: When subordinate clauses are nominalized, they lose their T-A-M marking. b. Stage II–re-finitization: Over time, for functional reasons (expressive power), nominalized clauses are gradually re-finitized and, among other adjustments, re-acquiring T-A-M marking.
Two things were wrong with hypothesis (31). First, the 1993 paper presented no time-continuum data to support the proposed two-stage scenario. So it may well be that nominalized subordinate clauses in Ute had never lost their T-A-M marking to begin with. And second, the 1993 paper presented no quantified distributional data, only hand-picked examples of T-A-M marking in nominalized subordinate clauses. So it may well be that those examples were possible but not statistically representative. While a quantified assessment of the distribution of T-A-M marking in Ute subordinate clauses is not quite the requisite data- base, it may help shed some light on a related question:
6 Ute, alas, is at neither point.
80 • What type of nominalized subordinate clauses are more likely to exhibit a wider range of T-A-M marking, and why?
Table (32) below presents the frequency distribution of T-A- M marking in subordinate clauses in three Ute oral texts: Two narratives about the distant past, where the bulk of the main-clause information is marked with the remote-past suffix -pga; and an expository discussion of current issues, with most finite main clauses marked by either the immediate/present suffix -y or the habitual suffix -mi(ya)7.
(32) Distribution of T-A-M in subordinate clauses in Ute oral texts subordinate-clause type ======ADV-clause OBJ-REL-clause SUBJ-REL-clause ======T-A-M suffix -ga -ku -chi -na -p -chi -t ======Mollie C. ======zero 3 6 1 1 1 1 3 HAB (mi) 1 1 / / / / / ANT (qa) / 11 / / / / / REM (pga) / 1 / 3 1 / / IRR (vaa) / / 1 / / / /
7 From Givón (ed. 2013). Text 1: “Sinawav and the Seven Sisters”, told by Mollie B. Cloud. Text 2: “The Last War Party”, told by Harry Richards. Text 3: “Talk to the Tri-Ute Language Conference” (1977) by Harvey Natchez. The texts are ca. 15 pp.-long each (three-line format).
81 Harry R. ======zero 6 12 3 2 / 1 / HAB (miya) / 2 / / / / / ANT (qa) / 1 / 1 / / / REM (pga) / 2 / / / / / IRR (vaa) / 1 / / / / /
Harvey N. ======zero 4 12 / 11 1 1 4 HAB (mi) / / / 2 / / / ANT (qa) / 6 / 7 1 / / REM (pga) / / / / / / / IRR (vaa) / 1 / 1 / / / ======TOTAL: 14 55 5 28 4 3 7
Examples of these subordinate clauses from the three texts are given in the Appendix, below. Two types of nominalized subordinate clauses were not included in the count: the SS complements of modal-aspectual-modal verbs (‘want’), and the DS complements of manipulation verbs (‘tell’). Their T-A-M marking is obligatory and limited to the irrealis suffix -vaa. The first thing to note in table (32) is that obligatory equi-subject subordinate clauses—the participial with -ga, the ADV-clause with -chi and the subject-REL-clauses with either -chi or -t—most often receive zero T-A-M marking. Put another way, referential continuity goes hand-in-hand with T-A-M continuity. The second observation is that even in the two clause-types where more referential freedom is in principle possible, ADV- clauses marked by -ku and object REL-clauses marked by -na, zero T-A-M marking is still very frequent, often with just the ante-
82 rior—here pluperfect marker— -qa. This once again suggests that the option of zero T-A-M marking is exercised judiciously, most likely in conjunction with referential continuity. Lastly, the distributions in (32) suggest that, as noted earlier, T-A-M marking in subordinate clauses, and thus their degree of finiteness, is controlled by general functional considerations that are largely independent of nominalization per se8.
6. In search of mechanisms
The recruitment of nominalization to create subordinate clauses is a natural, widespread diachronic phenomenon (see chs 25, 26 above). But as natural and widespread as it may be, the end result is often less than natural, tampering with the speakers’ normal/ frequent strategy for processing nominal case-marking and verbal morphology in main clauses. It may thus not be an accident that in all language families with historically-nominalized subordinate clauses, an eventual drift toward re-finitization is observed. In Uto-Aztecan, the conservative north (Numic, Takic, Yaqui, Huichol) shows little evidence of such a drift, yet. In the innovative south, on the other hand, re-finitization of subordinate clauses has been completed and nominalization has disappeared with few tra- ces. It is the middle strip—Guarijío, Tarahumara, perhaps Cora— that is of interest for studying the diachronic process of re- finitization9.
8 It is perhaps not entirely an accident that the name for ‘God’ in Ute, a lexical nominalization par excellence, is Núu-maroghoma-pga-t ‘He who created the people’, a nominalization that is still marked by the finite remote- past tense suffix -pga. 9 In So. Tepehuan, where all subordinate clauses have been re-finitized, the nominalizer suffix -d-am is still preserved, a relic of the old nominalized
83 In Guarijio, one finds several surviving old nominalizing suffixes in subordinate clauses, re-interpreted now as just part and parcel of the new finite morphology. The genitive marking of subjects in both object REL-clauses and ADV-clauses has disappeared in nouns, but it still survives in pronouns. Thus com- pare (Félix-Armandáriz 2006):
(33) a. Subject REL-clause: tihoé tapaná u’má-ka-(a)me man/S yesterday run-PAR-NOM ‘the man who ran away yesterday...’ b. Object REL-clause, nominal subject: kari amó karí-ta-ri-a Huaní house you/O house-build-PFV-NOM John/S ‘the house that John built for you...’ (hist.: ‘the house of John’s building for you’) c. Object REL-clause, pronominal subject: owítiame Mochibámpo no’ó tetewá-ri-a woman Mochibámpo 1s/G see-PFV-NOM ‘the woman I saw in Mochibámpo...’ (hist.: ‘the woman of my seeing in Mochibampo’) d. ADV-clause, nominal subject: temé neipá asi-má asi-só Huanita 2p/S last arrive-FUT arrive-NOM Juanita/S ‘we will arrive after Juanita has arrived’ (Hist: ‘we will arrive after Juanita’s arrival’) e. ADV-clause, pronominal subject: neipá yau-má-ni-a amó yau-só-pa last dance-FUT-1s-EMPH 2s/G dance-NOM-INCH ‘I will dance after you dance’ (Hist.: ‘I will dance after your dancing’) structure, in both subject REL-clauses and subject/agent lexical nominalization (García-Salido 2014). This suffix is probably cognate to the Guarijío/Tarahumara -ame subject REL-clause suffix and the Yaqui -me.
84 Complements of utterance verbs have been further re-finitized, with both their nominal and their pronominal subjects now marked as nominative. This may be due to analogical pressure from direct- quote complements, probably the most common complements for utterance verbs. Further, no nominalizing suffix has survived in such complements , although the participial suffix may be considered quasi-nominal. In contrast, the complements of cognition verbs such as ‘know’ still preserve a genitive pronominal subject and an old nominalizing suffix. Thus compare:
(34)a. Complement of ‘say’, nominative pronominal subject: apoé chaní temé noka-ri-áta wewe-ka 3s/S say 1p/S move-PFV-QU hit-PAR ‘he said that we hit him’ (hist. ‘he told our hitting him’) b. Complement of ‘know’, nominative nominal subject: aapóe nané-na peniátiame wikaht-ó María 3s/S know-PRES pretty sing-NOM Maria/S ‘They know that María sings pretty’ (hist.: ‘They know of María’s singing pretty’) c. Complement of ‘know’, genitive pronominal subject: nané-na-ne amó peniási-ka amó yau-yo know-PRES-1s 2s/GEN pretty-PAR 2s/GEN dance-NOM ‘I know that you dance pretty’ (hist.: ‘I know of your dancing pretty’)
Finally, both equi-subject and switch-subject verbal complements still take the old nominalizing suffix -(a)me, as in:
(35) a. Equi-subject complement of modal-aspectual verbs: simi-nare-ne ehtudiarwa-ni-áme kechewéka go-DESID-1s study-PRES-NOM Quechehueca ‘I want to study (the) Quechehueca (language)’ (hist.: ‘I want studying Quchehueca’)
85 b. Switch-subject complement of manipulation verbs: Hustína nahkí ki-kio’ko-ri-áme ini-míchio kuitá Agustina want NEG-get.sick-PFV-NOM be-PURP child ‘Agustina wants her child to be healthy’ (hist.: ‘Agustina wants her child being healthy’
Re-finization, the diachronic reversion to finite structure, it seems, may be gradual and piecemeal, progressing construction by construction and creating various intermediate structures, some more nominalized, others more finite. From the perspective of ideal iconicity and speech processing, such intermediate synchronic states make grammars less transparent, harder to learn and process, and in that sense ‘less natural’.
7. Discussion
The diachronic trajectory of erstwhile nominalized clauses may involve both the re-finitization of subordinate clauses and their de- subordination. Still, the two processes need to be considered independently, all along conceding that the facts relevant to re- finitization are rather meager, especially in terms of frequency distribution data. The suggestion made below are thus at best tentative.
7.1. Re-finitization of subordinate clauses
When a language re-finitizes its erstwhile non-finite subordinate clauses, chances are this is done gradually along well-known implicational hierarchies, the first of which is attested in Guarijío and probably has much to do with usage frequency and Zip’s law:
(36) nouns > pronouns
86 The other hierarchies probably have more to do with how independent a subordinate-clause is from its main clause in terms of the two main strands of thematic coherence—reference and tense-aspect-modality. This is where the Ute frequency distribution figures in table (32) may be relevant to our discussion, and where they seem to cohere with the—admittedly meager—Guarijio and Ute data:
(37) a. REL-clauses > V-complements b. Non-equi complements > equi complements c. Utterance V-complements > perception/cognition V-complements d. non-equi-subject ADV-clauses > equi-subject ADV-clauses
The vestige of old nominalizations that seems to survive longest, to judge by both the Tibeto-Burman (DeLancey 2011) and Uto- Aztecan data (Félix-Armandáriz 2006; García-Salido 2014), are the nominalizing verb suffixes, probably for the simple reason that they are quickly shorn of any meaning, and are soon re- interpreted as just part of the tense-aspect marking of finite clauses. This is just as true of the Ute remote-past marker -pga (see (23) above) as it is of the English progressive marker -ing. This is, I suspect, the story of survival of the Uto-Aztecan subject nominalizer -(a)me along the cline of re-finitized subordinate-clause syntax:
Yaqui > Guarijío > So. Tepehuan
7.2. De-subordination and vestigial nominal structure in main clauses
The story of how vestigial traces of nominalized morpho-syntax survive in main clauses is the story of de-subordination and
87 subsequent re-structuring. In principle, this story has little to do with nominalization per se. However, if a subordinate clause happens to have been nominalized in its former incarnation, its nominal features will surface out in main clauses as by-products of de-subordinated. This is how the ergative morphology in Tibeto-Burman and Cariban main clauses came into being, as by-product of de- subordination (Gildea 1998; DeLancey 2011). This is how the Ute nominalizer -p became part of the remote-past marker -p-ga in main clauses. This is how the Japanese genitive suffix -ga was re-analyzed as a nominative marker (Akiba 1978; Shibatani 2007). This is also how the Numic and Takic genitive suffix -y/-i became an object suffix, as by-product of the de-subordination of nominalized VPs (see ch 18). And this is, lastly, how subordinate verb-complement morphology was pulled into subjunctive and hortative main-clauses in Bantu (Givón 1971b) and Ute (see (24) above). De-subordination is a highly universal mechanism, transcending the typological distinction between nominalizing and finite languages. When it occurs in a nominalizing language, through whatever mechanism, de-subordination introduces the nominalized structure of erstwhile subordinate clauses into main-clause syntax.
ABBREVIATIONS
ANT anterior 1s 1st person singular DESID desiderative 2s 2nd person singular DIR directional 3s 3rd person singular DS different subject 1p 1st person plural EMPH emphasis 3p 3rd person pliural FUT future du dual G genitive HAB habitual
88 IMM immediate INCH inchoative INCL inclusive IO indirect object IRR irrealis LOC locative NEG negative NOM nominalizer O object PAR participle PASS passive PFV perfective PL plural PRES present PURP purposive QU quotative REM remote S subject SS same subject T-A-M tense-aspect-modality
89 References
Akiba, K. (1978) Topics in Japanese Historical Syntax, University of California, Los Angeles, PhD dissertation (ms.). Austin, P. (1992) “Cases and clauses in Jiwarli, Western Ausrtralia”, La Trobe University Working Papers in Linguistics, 5. Bisang, W. (2011) “Finiteness and nominalization—convergence and divergence”, Symposium on Finiteness and Nominalization, CNRS, Paris (ms.). Bommelyn, L. and T. Givón (1998) “Internal reconstruction in the Tolowa Athabaskan verb”, Cuarto Encuentro Internacio- nal de Linguística en el Noroeste, Universidad de Sonora, Hermosillo, Son. Carlson, R. (1994) A Grammar of Supyire, Berlin/NY: Mouton De Gruyter. DeLancey, S. (2011) “Finite structure from clausal nominalization in Tibeto-Burman”, in Yap, F.-H., K. Grunow-Harsta and J. Wrona (eds.), Nominalization in Asian Languages: Diachronic and Typological Perspectives, 343-62. Amsterdam: J. Benjamins. Dixon, R.M.W (1991) A New Approach to English Grammar Based on Semantic Principles, Oxford: Oxford University Press. Evans, N. (2007) “Insubordination and its uses”, in I. Nikolaeva (ed.) Finiteness: Theoretical and Empirical Foundations, Oxford: Oxford University Press. Félix-Armandariz, R. (2006) A Grammar of River Warihío, Rice University, Houston, TX, PhD. Dissertation (ms.). García-Salido, G. (2014) Clause Linkage in Southeastern Tepehuan, a Uto-Aztecan language of Northern Mexico, University of Texas, Austin, PhD. Dissertation (ms.). Genetti, C., A.R. Coupe, E. Bartee, K. Hildebrandt and Y.-J. Lin (2008) “Syntactic Aspects of Nominalization in five Tibeto-
90 Burman languages of the Himalayan Area”, Linguistics of the Tibeto-Burman area, 31.2. Gildea, S. (1998) On Reconstructing Grammar, Oxford: Oxford University Press. Givón, T. (1971b) “Dependent modals, performatives, factivity, Bantu subjunctives and what not”, Studies in African Linguistics, 2.1. ______. (1975) “Serial verbs and syntactic change: Niger-Con- go”, in C. Li (ed. 1975). ______. (1983) “Topic continuity in discourse: An introduction”, in T. Givón (ed. 1983). ______. (ed 1983) Topic Continuity in Discourse: Quantitative Cross-Language Studies, TSL #3, Amsterdam: J. Benjamins ______. (1993) “Nominalized clauses in Ute: The diachronic seesaw of finite structure”, 2do Encuentro Internacional de Linguística en el Noroeste, Universidad de Sonora, Hermosillo, Son. ______. (2001) Syntax: An Introduction (2 vols), Amsterdam: J. Benjamins. ______. (2009) The Genesis of Syntactic Complexity, Amsterdam: J. Benjamins. ______. (2011) Ute Reference Grammar, Amsterdam: J. Benjamins. ______. ed. (2013) Ute Texts, Amsterdam: J. Benjamins. Haiman, J. (1983) “On some origins of switch reference marking”, in J. Haiman and P. Munro (eds), Switch Reference and Uni- versal Grammar, TSL #2, Amsterdam: J. Benjamins. Hill, J. (2005) A Grammar of Cupeño, Berkeley and Los Ungeles: University of California Press. Hopper, P. and S. Thompson (1984) “The discourse basis for lexical categories in universal grammar”, Language, 60. Hyslop, G. (2011) A Grammar of Kurtöp, PhD. Dissertation, University of Oregon, Eugene (ms).
91 Lewis, G.L. (1967) Turkish Grammar, Oxford: Clarendon Press. Li, C.N. (ed. 1975) Word Order and Word Order Change, Austin: University of Texas Press. Matisoff, J. (1972) A Grammar of Lahu, Berkeley: U.C. Press. Mithun, M. (1991) “The role of motivation in the emergence of grammatical categories: The grammaticalization of subjects”, in E. Traugott and B. Heine (eds 1991). ______. (2011) “Interaction between nominalization and finiteness”, Conference on Nominalization and Finiteness, CNRS, Paris, Sept. 2011. Osam, E. K. (1994) Aspects of Akan Grammar, PhD. Dissertation, University of Oregon, Eugene (ms). Shibatani, M. (2007) “Relativization i n Sasak and Sumbawa, eastern Indonesia, with some comments on Japanese”, Semi- nario de Complejidad Sintáctica, Universidad de Sonora, Hermosillo, (ms.). Thurman, R. (1978) Interclausal Relations in Chuave, University of California, Los Angeles, MA thesis (ms). Traugott, E. and B. Heine (eds 1991) Approaches to Gramma- ticalization, TSL #19 (2 vols), Amsterdam: J. Benjamins. Watters, D. (1998) The Kham Language of West-Central Nepal (Takale Dialect), PhD. Dissertation, Eugene: University of Oregon (ms.). ______. (2008) “Nominalization in Kiranti and Central Himalayan languages”, Linguistics of the Tibeto-Burman area, 31.2. Weber, D. (1996) Una Gramatica del Quechua del Huallaga, Seria Linguística Peruana 40, Lima, Perú: Instituto Linguístico de Verano. Zipf, G. (1935) The Psychobiology of Languages, Cambridge, MA: MIT Press.
92 APPENDIX: Examples of T-A-M marking in Ute nominalized subordinate clauses
Tex #1: Mollie Cloud, “Sinawav and the seven sisters”
clause type T-A-M marking ======1. ...yúaa-va-t-am tavi-navichi-ku... flats-at-DIR-3p step-MASS-SUB ‘...when they landed in open country...’ ADV-cl, -ku, zero
2. ...kh-’ura-’uru ‘uwayas p-pi-kyay-ku-’uru... then-be-that 3s/G RED-sleep-ANT-SUB-that ‘... then after he slept...’ ADV-cl, -ku anterior -qa
3. ...tna-khwa-pgay-ku-’ura-’uru hunt-go-REM-SUB-be-that ‘...so as he was going hunting...’ ADV-cl, -ku remote -pga
4. ...’uru-’ura nnay ya’ay-kwa-qhay-ku-n that-be 1s/G die-go-ANT-SUB-1s ‘...when I (will) have died...’ ADV-cl, -ku anterior -qa
5. ...wcha-rp’nap ‘uru tka-miya-ku... calf-muscle/O that/O eat-HAB-SUB ‘...when you eat calf muscles...’ ADV-cl, -ku habitual -mi(ya)
6. ...kh-’ura míyh-t-aa-s tavi’wa-gha then-be far-NOM-O-C step-PAR ‘...then stepping farther...’ PAR-cl, -ga zero
93 7. ...kh-’ura ‘um súuva-t-m-aa-ni ‘um then-be 3p/O other-NOM-PL-O-like 3p/O ‘...and he told all those others nana-chigya-qha-na-av ‘áy-pga REC-race-ANT-SUB-OWN say-REM that he was going to race...’ V-COMP -na anterior -qa
8. ...púupa máy-pga-na-’u way say-REM-REL-3s ‘...the way he said (it)...’ O-REL-cl, -na remote -pga
9. ...pachichi-’u piwa-ri-vaa-na sister-3s marry-do-IRR-REL ‘...the one his sister was supposed to marry...’ O-REL-cl, -na irrealis -vaa
10. ...p-paa-t-am káaya-na-p REL-DIR-NOM-3p stash-REL-NOM ‘...where they used to stash food...’ O-REL cl, -na zero
11. ...nnay pachichi-aa-n ‘uni-pgaa-p ‘uru 1s/G sister-G-1s do-REM-NOM that/O ‘...the one my sister did that to...’ O-REL cl, -p remote -pga
12. ...‘ums-’ura ‘ura-’ay Sinawavi nghwa-p 3p-be be-IMM Sinawav/O bear-NOM ‘...it was them that she bore (to) Sinawav...’ O-REL cl, -p zero
13. ...tuku-sa’map-ga-t ‘ura-vachi... cougar-blanket-have-NOM be-IRR-BG ‘...it was the one who had the cougar blanket...’ S-REL-cl, -t zero
94 14. ...”sh-sh-sh” máy-chi-’uru... sh-sh-sh say-NOM-that ‘...the one making the “she-she-sh” sound) S-REL-cl, -chi zero
Text #2: Harry Richards “The last war party”
1. ...má-vaa-na-uv pagha-’ni-mi kani-gyay-ku-n there-at-LOC-TOP go-INT-HAB house-have-SUB-1s ‘...he used to live there when I had my house (there)...’ ADV-cl, -ku zero
2. ...kukwachi ‘uway ‘i-vee-k pagha-’ni-miya-ku... Mexican/G that/G here-at-EM go-INT-HAB-SUB ‘...when that Mexican used to live here...’ ADV-cl, -ku habitual -miya
3. ...kh-’ura-’uru wíichk-vaa-ku then-be-that morrow-IRR-SUB ‘...then when it would become morning...’ ADV-cl, -ku irrealis -vaa
4. ...’ú-vway-aqh-’ura-’uru núuchi-u s-sti’i-pgay-ku, there-at-it-be-that Ute-PL RED-feel-REM-SUB ‘...then when the Utes discovered (it)...’ ADV-cl, -ku remote -pga
5. ...psariniya-qha-qay-ku-’uru... story.tell-PL-ANT-SUB-that ‘...when they told the story...’ ADV-cl, -ku anterior -qa
6. ...ma-vaa kwáa tgaay’wa-chi... there-at INTJ arrive-NOM ‘...and upon arriving there...’ ADV-cl, -chi zero
95 7. ...ku-kukwi-gya qo-qho’ay-paa-ni-am... RED-shoot-PAR RED-slaughter-IRR-FUT-3p ‘... (thus) shooting we’ll slaughter them...’ PAR-cl, -ga zero
8. ...’an-’apagha-qa-na... just-speak-ANT-REM ‘...in case I spoke to them...’ O-REL-cl, -na anterior -qa
9. ...p-paa-y nagukwi-kya-na-p... REL-DIR-O fight-PL-REL-NOM ‘...the place where they fought...’ O-REL-cl, -na zero
10. ...‘ums-’uru pá-ini máy-chi-m... 3p/S-that three-S say-NOM-PL ‘...those three who were planning (the ambush)...’ S-REL-cl, -chi zero
Text #3: Harvey Natchez “Talk to the Tri-Ute language conference”
1. ...’i-vee-naagha-’uru tga’wi-ku... here-at-in-that arrive-SUB ‘...when it arrives here...’ ADV-cl, -ku zero
2. ...’uni-paa-chi-ku-’uru... do-IRR-NOM-SUB-that ‘...when they want to do it...’ ADV-cl, -ku irrealis -vaa
96 3. ...’an-a togo[y]-ni-’ura máy-kya-qhay-v WH/QU-O well-like-be say-PL-ANT-NOM ‘áy-kya-ni-’uru say-ANT-like-that ‘...I was wondering how well they were talking...’ V-COMP, -v anterior -qa
4. ...’ichay-’uru máy-kya-na-av... this/O-that say-PL-REL-OWN ‘...this thing that they say...’ O-REL-cl, -na zero
5. ...’ichay ‘iya-na ‘umas ‘áy-kya-qha-na this/O here-LOC 3p/G say-PL-ANT-REL (this is then what they have said) O-REL, cl. -na anterior -qa
6. ...mrkachi púupa máy-miya-na. Whiteman/G way say-HAB-REL ‘...the way the white man speaks...’ O-REL-cl -na habitual -miya
7. ...púupa-aqh-’uru ‘ura-vaa-na; way-it-that be-IRR-REL ‘...the way it should be...’ O-REL-cl, -na irrealis -vaa
8. ...mamas-’uru p-paa-tugwa-am-’uru máy-kya-p.. 3p/G-that REL-DIR-go-3p-that say-PL-NOM ‘...what they have been talking about...’ O-REL-cl -p zero
9. ...’ina-khwa-tna-khwa ‘uni-kya-’ni-chi-m. here-go-climb-go do-PL-INT-NOM-PL ‘...those who live up here...’ S-REL-cl, -chi zero
97 10. ...’ums education-i-vee-t ‘ura-qa-t-m, 3p/S education-O-at-NOM be-PL-NOM-PL ‘...those who are in the Education Dept. S-REL-cl, t, zero
98