1. Inverse Number and Noun Class
One of the most striking characteristics of the Kiowa-Tanoan family is inverse number, so-called, because the relevant morphemes “invert” lexical number. Consider ‘bugs,’ ‘sticks,’ and ‘hair’ in Kiowa. The basic forms of these nouns are numerically heterogeneous: respectively, pól is singular/dual, á dual/plural, ául just dual. Crosscutting this diversity, a single suffix delivers the missing numbers of all three: pó•-dáu ‘bug-inv’ is plural, á•-dau ‘tree-inv’ singular, and áu•-dáu ‘hair-inv’ nondual.
Inverse marking occurs not only on nouns, but on their agreeing modifiers, on relative clause suffixes, and on verbs (though, in Tewa, only for animates). For instance, in Kiowa, á• ‘sticks’ occurs with basic (bas) -de on demonstrative and relative marker and straightforward plural agreement gya- on matrix and subordinate verb.
By contrast, with inverse á•-dau ‘stick,’ the demonstrative and relative marker share inverse -gau and the verbs, inverse e-.
The same inverse markers occur when the inverse has plural reference, as in:
Basic forms can also be overtly marked. This is particular common in Tiwa, as the following basic/inverse pairs from Isletan (Southern) Tiwa (Tucker Harrington, 1920a) illustrate: t’a-’ide / t’á-mnin ‘deer-bas/inv,’ kar-íde / kár-nin ‘wolf-bas/inv,’ p’akhəłá-de / p’ak<001>łha-n ‘star-bas/inv.’ (The allomorphic variety in these examples is typical for the family as whole; the initial Kiowa examples were chosen to avoid this.)
Kiowa and the Tanoan languages differ in the treatment of dual. No Kiowa noun with dual-denotation is inverse marked (basic forms always include dual), and (excepting first person) none triggers inverse agreement. In Tanoan, nouns that take inverse marking do so in the dual, as in Jemez singular/dual p<002>•-š ‘drum-inv,’ dual/plural vê•læ-š ‘man-inv,’ and dual-only bél<003>-š ‘loaf-inv’ (Yumitani, 1998). Moreover, inverse-marked nouns can trigger dual or inverse agreement, depending on their denotation. In Isletan Tiwa (Tucker Harrington, 1920a, my glossing), for instance, dual/plural ’u’u-n ‘baby-inv’ agrees as 3d’in- for two babies (4), 3i’i- for more (5).
Despite these and other above-mentioned differences, the core concept of inverse marking applies across the family.
Moving beyond description, it is tempting to tackle the meaning of the inverse head on. A naive approach would, for instance, propose that it means ‘unexpected number of.’ But the Jemez (for whom vêê læ-š ‘man-inv’ is dual/plural) surely do not think that pluralities of people are unexpected; they are, after all, not hermits. Nor could the Kiowa (for whom basic ául ‘hair’ is dual) think that hairs come in pairs. This recalls Quine’s dictum that “Wanton translation can make natives sound as queer as one pleases.”
A more fruitful tack takes heed of the insight, common to descriptive and theoretical studies alike, that Kiowa-Tanoan nouns with inverse forms have an inherent number (like the gender specifications of more familiar languages) and that inverse occurs when there is a conflict between inherent number and denotational number. For instance, Kiowa á• ‘sticks’ is inherently dual/plural; so, conflict, and hence inverse á•-dau, occurs when denotational number is singular. To the extent that inverse has a meaning, it is, then, as a dependent of the meanings of inherent and denotational number. As denotational number poses general crosslinguistic questions, the specifically Kiowa-Tanoan issues are how to represent inverse and how to explicate the meaning of inherent number. A lightly theoretical answer is presented, so as to frame the topic of noun class.
Conflict can be captured just with features generally used for number (recourse to a sui generis [inverse] feature is otiose). Let singular be [+atomic +minimal] (atomic and, hence, minimal); dual, [–atomic +minimal] (minimally nonatomic); and plural, [–atomic –minimal] (nonminimally nonatomic). Then, for dual/plural á• ‘sticks,’ the inherent feature is that common to dual and plural, –atomic. When inherent –atomic meets a denotational singular [+atomic +minimal], the values of +atomic and –atomic conflict. The correct distribution of inverse-marking follows if inverse is the morphological response to +/– conflicts; hence the inverse-marked singular a•-dau ‘stick.’ (Similarly, if hair is inherently [–atomic +minimal], then inverse will arise for singular, in virtue of +atomic, and plural, in virtue of –minimal.) Approaches along these lines are developed in detail in Harbour (2007, 2011b), Noyer (1992), Watanabe (2015).
These representations transform the question of inherent number into what it means for ‘sticks’ to be inherently nonatomic, for singular/dual ‘bugs’ to be minimal, and hairs, both. The answer involves the semantic characteristics of the noun classes, of which ‘sticks,’ ‘bugs,’ and ‘hair’ are representatives.
Kiowa-Tanoan noun classes can be named mnemonically. For instance, Kiowa ‘bug’ belongs to the sdi class, because it takes singular (s) agreement for one bug, dual (d) agreement for two, and inverse (i) agreement for three or more:
Similarly, ‘stick’ belongs to the idp class, triggering i agreement in the singular, d agreement in the dual, and p agreement in the plural:
The sdi and idp classes are two of the most numerous in Kiowa. Featurally, sdi nouns are all +minimal, idp nouns –atomic. The most salient difference between them is animacy: all animates are sdi nouns, all idp nouns, inanimates. Additionally, there are sdi inanimates, covering things capable of independent motion (sun, moon, rivers; cars, wagons, wheels), implements of cutting (knives, scissors, hoes), and some mobile or cutting body parts (legs, tongues, necks; teeth). Typical idp nouns are plants and their parts (sticks, flowers, leaves), artefacts (buckets, paper, arrows), and other body parts (bones, feathers, feet).
The connection between animacy and minimality lies with motility and individuability (following Noyer, 1992). For humans and other large animals, this is straightforward: we move and are individuals. Smaller animals, even those that swarm, are also strongly individuable: faced with several bugs, one does not doubt where one ends and the next begins, as independence of motion makes clear where the bounds of each individual lie. Thus, motility forms the conceptual core of sdi class membership and provides the conceptual nexus to the class feature +minimal as independent motion makes the bounds of each minimal unit, or individual, clear.
Similar reasoning applies to the –atomic idp class. One might expect animacy and inanimacy to be coded by different values of the same feature, such as +minimal and –minimal. However, values are only one form of semantic complementarity that the number features provide. By definition, atoms are minimal, and anything nonminimal is also nonatomic. So, there are entailment relations from +atomic to +minimal and from –minimal to –atomic. As a result, +minimal and –atomic represent conceptual poles. A class with the opposite defining characteristics of motility and individuability is reasonably represented as the polar featural opposite of +minimal, hence, as –atomic.
This approach solves the mystery of ‘hair.’ Mnemonically, this is an idi noun: i agreement for one or many hairs, d agreement for just two. It is classmate with some other forms of body hair and with midsize fruit, like apples and plums. The current approach claims that idi nouns, which pool the class features of sdi and idp nouns, must have joint properties, with nexuses to both features. The joint properties of ‘hair’ are familiar from European languages: Italians wash their ‘hairs’ (capelli), English, their hair, and Germans, either (Haar or Haare). Hair’s undifferentiable plurality makes it –atomic, like idp leaves, but in forming a single, loosely mobile mass, it resembles a +minimal sdi noun. Similarly, apples, plums, and similar sized fruit grow in profusions typical of inanimates; however, their individual bounds are robust and salient (particularly as regards their use by people). The apparent inherent duality reduces to an epiphenomenon, arising from the encoding of two separate properties via the features that, jointly, represent dual number.
Nouns may furthermore simply eschew inherent number. These are simply sdp nouns, displaying transparently denotational agreement for all numbers, without inverse. This approach predicts members of the sdp class to display two negative semantic characteristics. They should not display the hallmark characteristics of other classes, such as animacy, nor should the class itself display strong internal semantic coherence. These predictions hold. The class, which is rather small, comprises one natural object (rock), one item of adornment (earring), some articles of footwear (moccasin, boot), and a few implements (rope, key).
Matters are similar for Tanoan. Noun class is best described for Jemez. Consolidating Sprott (1992) and Yumitani (1998), Harbour (2011b) shows that there is again an animate sdi class and an inanimate idp class, but, given the tendency of dual agreement to syncretise with inverse agreement, and for dual-denoting nouns to bear inverse marking, these classes might be renamed sii and iip. In a similar vein, there is a very small iii class, which, like Kiowa idi, contains nouns that have properties of weak individuability, and an sip class, paralleling Kiowa sdp, without strong internal properties. Descriptions of noun classes in other Tanoan languages are less thorough, but one can discern the same system at work.
Though this leaves much unsaid about other noun classes (for collective, mass, and pluralia tantum nouns), the foregoing covers the essentials of this quintessentially Kiowa-Tanoan phenomenon. Sutton (2010) provides a useful starting point for further comparison.
A second hallmark of the Kiowa-Tanoan family is what Harrington (1916) called “so elaborate and pretentious a plan” of agreement. The Kiowa-Tanoan agreement prefix registers up to three arguments within a unit at times so tightly fused that disaggregating constituent morphemes is a challenge. For instance, in Kiowa â•- ąu• ‘they gave it to him,’ the verb is simply <004>u•, meaning that â•* (‘*’ lowers verb tone) encodes a third plural agent and third singular applicative and object (3a:3s:3s). For reasons of space, only the Kiowa and Taos systems are discussed in detail here. These serve as a guide to the rest of the family, which receive briefer treatment.
The full Kiowa system is presented in table 1 (following Harbour, 2007). The basic division is between transitive prefixes with third person objects (upper portion) and ditransitive prefixes (lower). On the basis of morphological affinities, plain intransitive prefixes are included in the upper portion, and intransitives with applicatives as well as transitives with nonthird person objects are presented in the lower part.
To illustrate, consider the top 1s row. Gya, the form in the :3s column, encodes 1s:3s, a first singular agent and third singular object. Subsequent columns show plural, dual, inverse, and animate plural / reflexive agreement (discussed shortly). Hence, 1s:3p, first singular on third plural, is gyat; 1s:3d is nen, 1s:3idé, and 1s:3a/rxde. The leftmost column, without a heading, gives intransitive agreement, 1sa. (The 3p line, last of upper portion of the table, exists only as intransitive agreement: referencing inanimates, it cannot occur as the agent of transitives, which are restricted to animates.)
For the lower portion, consider the third (1s):2s row. This gives some 2s applicatives and related prefixes. For example, em, beneath the intransitives of the upper portion, encodes transitive 1s:2s. All other columns give ditransitives and intransitive-cum-applicatives. For instance, gyá means both 1s:2s:3s (‘I verb it for you’) and :2s:3s (‘it verbs for you’).
Many rows in the lower portion contain bracketed agents. This reflects the fact that null exponence is frequent, particularly for agents of ditransitives with nonsingular applicatives. This results in ambiguity between ditransitive and intransitive-cum-applicative prefixes. For instance, the third dual possessor of third inverse prefix :3d:3imén is also used for any agent acting on a third inverse object for third dual, 1/2/3:3d:3i.
Table 1. Kiowa Agreement Prefixes.
Despite the substantial quantity of null exponence, several prefixes directly encode all arguments. Triple encoding occurs for singular applicatives, even if the applicative in these cases is only partially represented. This is most obvious for 1s/3s applicatives (last rows of table), all of which end in falling tone and cause low tone on the verb (*). Consider 2d:1s/3s:3dménê•*. This is clearly comparable to the transitive prefix 2d:3dmén, which explicitly encodes agent (m-) and object agreement (-én). The additional falling vowel of ménê•* is thus attributable to a third argument, the applicative.
A second instance of triple encoding, in the 1s:3s row, is robust across the family. It can be seen by considering the morphological increment to ditransitive from either transitive or intransitive-cum-application. Take transitive 1s:3sgya. The applicative results in additional exponence, 1s:3s:3sgyá, via the high tone that is present for all ditransitive prefixes. Thus gyá encodes agent and object, like gya, while also registering the presence of an applicative. (The same relationship is transparent between 1s:3dnen and 1s:3s:3dnén; plural and inverse objects are opacified by allomorphy.) Alternatively, the third singular intransitive-cum-applicative prefixes also clearly register two arguments. For instance, :3s:3iáu uses a vowel that occurs for inverse only in prefixes with applicative agreement. To this, a first singular agent adds the exponent g; hence the triple encoding 1s:3s:3igáu. (Other prefixes of this row decompose the same way, 1s:3s:3sgy-á, 1s:3s:3py-án, 1s:3s:3dn-én, noting the vowel- and nasal-dependent allophony of g, g~y~n~d).
As just hinted, tone bears a substantial grammatical burden. Prefixes can have high, low, or falling tone, and can impact the tone of the subsequent verb. In fact, a has four different tonal forms, all with different meanings: 1sa, 3aá, 3a:3sá*, 3a:1s/3s:3sâ•* (length is excrescent of falling tone). The tone lowerer is unique to Kiowa, but such tonal intricacies are shared by all members of the family, except Jemez, which has effaced all tone from agreement.
Besides representing the general pattern of agreement systems across the family, Kiowa is an interesting starting point because of some additional complexities.
First, it has more agreement categories than the Tanoan languages. On the one hand, it displays a clusivity contrast. The forms in question are not inclusive or exclusive per se, but are syncretic: inclusive nonsingular with second inverse, and exclusive nonsingular with third inverse. Paucity of morphological means notwithstanding, Kiowa is the only language of the family to display clusivity in agreement. (Kiowa pronouns themselves do not mark the distinction. By contrast, Jemez has a clusivity distinction in its pronouns but not, it appears, in its agreement. Other languages have only a general first person.)
Second, Kiowa has five third person categories. Besides singular, dual, plural, and inverse, there is a category for third person plural animates (3a), normally those with whom the speaker has empathy, such as fellow tribe members. As a direct object, 3a agreement is syncretic with reflexives. So, de-hól is ambiguous between 1s:rx ‘I killed myself’ and 1s:3a ‘I killed them (higher animates).’ (When there is overt applicative agreement, reflexives and animate plurals objects syncretize with dual or inverse.)
An important feature of Kiowa (and Tewa) is that every argument combination is encoded by some prefix (person restrictions on objects in the presence of applicatives aside). Often, morphological absence of the agent or applicative results in homophonous agreement for partially distinct argument combinations. However, whether by non-, partial, or complete encoding, every argument combination is mapped onto a prefix. This is not the case for the Tiwan languages (or Jemez), to which we now turn.
Table 2 presents the agreement system of Taos (Northern Tiwa). It is historically noteworthy as the first to have been extensively and accurately described, in an obscure (1916) publication by the redoubtable J. P. Harrington. The version presented below, which marks tone (something that Harrington did not consistently do for any Kiowa-Tanoan language until the 1940s), is from C. Kontak and J. Kunkel (from their unpublished work, “Grammar sketch of Northern Tiwa, Taos dialect,” no author location or affiliation given; first draft 1986, revised 1987). It is strongly parallel to the Southern Tiwa system as presented by Allen, Frantz, Gardiner, and Perlmutter (1990), but there, again, complete tone information is lacking. The system presents a range of pairwise syncretisms, between intransitives and :no (explained below), :no and :3s, and :3s and :3i, depending on the person and number of the agent and applicative; attention should be paid to alignment with column headings and braces in reading the table.
Taos differs crucially from Kiowa in banning some argument combinations: third persons cannot act on first or second, whether in transitives (*3:1, *3:2) or ditransitives (*3:1:3, *3:2:3). Rather, such combinations demand passivisation of the verb. This phenomenon has been extensively documented for Southern Tiwa, from which the following examples are drawn (Allen et al., 1990). Passivized transitives agree with the underlying first or second person object, as in (8).
Passivized ditransitives, like (9), show no object advancement and use the intransitive-cum-applicative agreement of (10).
These restrictions are represented in the final row of Table 2, in which every cell is starred for ungrammaticality. Allen et al. (1990) report that the Southern Tiwa ban on third person agents extends to action on applicatives of any person, including *3:3:3, but Tucker Harrington (1920a, 1920b) does not seem to support the extended ban. Such 3:3:3 combinations are licit in Taos.
These restrictions aside, Taos and Kiowa make the same broad divisions. In the upper portion of the table are, again, intransitive (left) and transitive prefixes; in the lower portion, applicatives and transitives with first and second person objects. The columns, though, present two major differences from Kiowa. First, reflexives are only reflexive: plural animate 3a does not exist and plural animates are 3i. Second, there is no :3d column. Northern and Southern Tiwa have wholly lost dual object agreement (in Jemez, it is vestigial, restricted to 1/2/3.s:3d). Instead, singular, inverse, or plural agreement is used, depending on class. An addition, though, is the :no column. This number-invariant form is used for nouns that trigger occurrence of no directly after the agreement prefix, as in ti-nó-éu-mu ‘I saw a leaf/leaves’ (1s:no-no-leaf-see). Its forms are generally syncretic with 3s object agreement.
Table 2. Taos Agreement Prefixes.
Like Kiowa, Taos presents tonally minimal pairs. For instance, for nonsingulars, intransitive and transitive :3s forms are distinguished by low/high tone, as in 2dmon and 2d:3smón (cf, Kiowa 2dma and 2d:3smá*). 2d, in fact, offers a tonally minimal triple, via its falling tone reflexive, 2d:rxmôn.
Taos, also like Kiowa, presents cases of triple encoding. For instance, 2i:3i is mopí, with mo- attributable to the second person (2imo) and -pí to the inverse object (compare 3ii, 3i:3iipí). This prefix serves as the base for 2i:3:z combinations, such as 2i:3:3s/imopîm and 2i:3:3pmopîw, which thus encode agent, applicative, and object. Paralleling Kiowa, 1s:3s:z ditransitives decompose into agentive and applicative-cum-object parts. Alongside :3s:3sǫ, :3s:3ióm, and :3s:3pów, there are 1s:3s:3stǫ, 1s:3s:3itóm, and 1s:3s:3ptów (where first singular t is recognizable from transitive 1s:3sti and reflexive 1s:rxtǫ.)
As stated above, the Southern Tiwa system strongly resembles that of Northern Tiwa. By contrast, the remaining two systems of the family, those of Jemez and Tewa, have undergone substantial changes. In Jemez, these are mostly simplifications, resulting in much greater morphological transparency and in many syncretisms of s/p and d/i object agreement. In Tewa, the changes are more radical, and the system exhibits substantial innovations. These include use of applicative agreement for animate direct objects (which preserve number marking), loss of number marking for inanimate objects, and development of imperative prefixes.
So brief an overview of such complex systems must, naturally, leave many matters unmentioned. However, the examples chosen highlight the range of variation that the family as a whole presents, while underlining unifying factors, such as underlying organization, the pervasiveness of triple encoding, the systematic relationship between :3s:3 and 1s:3s:3 prefixes, and the importance of tone.
3. Direct Object Incorporation
Amongst syntacticians, the Kiowa-Tanoan languages are best known for incorporation (and the obligatory passives touched on above), thanks to a series of papers by Allen, Frantz, and Gardiner, which culminated (but did not end) with Allen, Frantz, Gardiner-Perlmutter (1990), and fed into influential works by Rosen (1990) and Baker (1996). Incorporation is present in Southern Tiwa (here, Istelan) (11), Northern Tiwa (here, Picurís) (12), and Tewa (13):
It is however absent from Kiowa (15):
Thus, in contrast to inverse marking, noun class, and super-rich fusional agreement, incorporation, like obligatory passive (and agreement restrictions), is confined to a subpart of the family. Tiwa and Jemez display both. Tewa has incorporation without obligatory passive. Kiowa has neither.
Morphologically, incorporated nouns differ from free ones in lacking basic and inverse suffixes. For instance, Southern Tiwa kan-ide ‘horse’ (horse-bas) and kan-in ‘horses’ (horse-inv) are identical as incorporates, kan-:
Number of the incorporate is still discernible from the agreement prefix, men for singular, mimim for inverse. This persistence of object agreement reveals that Tiwa (and Tewa) incorporation is essentially different from that in other languages (Baker, 1996) in not reducing the valence of the predicate: mų ‘see’ in (16) takes transitive agreement prefixes.
Absence of basic/inverse suffixes shows that incorporates are smaller than free nouns. However, (12) shows that incorporation is not restricted to heads in Picurís: łoła-phon (butterfly-black) and łoła-ts’ol (butterfly-yellow) are incorporated nouns modified by adjectives. The same holds for Tewa (17) and Southern Tiwa (18), which illustrate incorporation of noun-noun compounds, the second of which even bears a diminutive suffix:
The discourse function of incorporation also sets Tewa and Tiwa apart from languages classically described as incorporating. Illustrating with examples from a short Picurís story (Harrington & Roberts, 1928, my glossing), incorporation can introduce a new entity, like kāl- ‘wolf’ in (19), that is subsequently resumed by a free noun phrase, like kāl-enę (wolf-bas) in (20):
Conversely, a previously established discourse entity, like the woman in examples (19)–(20), can move from free noun in (19) to incorporate in (21) and back again in (22).
There is therefore no requirement that incorporated nouns be indefinite, nor are they restricted to nonreferential uses. Rather, they have all the referential and discourse possibilities typically available to free nouns.
Despite their similarities, Tewa and Tiwa incorporation differ from each other in two regards. First, incorporation is obligatory in 3:3 transitives in Tiwa and for objects of verbs with applicative arguments.
There are no circumstances under which incorporation is obligatory in Tewa, to judge from the corpus study, Harbour, 2013. Instead, one finds incorporated/free pairs, like (13) and (24):
Second, Tiwa incorporation create discontinuous constituents, with modifiers external to the verb and the noun itself incorporated (with any modifiers):
If such constructions are possible in Tewa, the corpus does not show them. Instead, quantified nouns are always free, as exemplified by wí píví ‘some meat’ in (24).
The differences that incorporation displays in Tewa and Tiwa (and other languages) raise the possibility of fruitful comparative work, leading to a deeper understanding of the parameters of incorporation. However, the somewhat neglected state of Tewa linguistics means that such work has yet to be undertaken.
The Kiowa-Tanoan languages display a fascinating array of unusual grammatical features. Inverse number, noun classification, super-rich fusional agreement, agreement restrictions, and morphologically burdened tone are characteristic across the family (incorporation, less so). However, rarely does a characteristic present identically in any two members. This holds equally for many other topics that have been excluded for reasons of space or because they are insufficiently studied (e.g., mass nouns, suppletion, negation, tense versus aspect, evidentiality, word order, relative clause structure, and headedness). Given the endangered state of much, if not all, of the family, the opportunity for future fieldwork may be limited, but the existing body of documentation, though scattered in obscure loci, is rich enough to support much further research.
The reader is referred, initially, to Harbour (2011a) and Sutton (2014) for overviews. Thereafter, for pure grammatical description, the best single grammar of a Kiowa-Tanoan language is Watkins’ (1984) description of Kiowa, which synthesizes much earlier literature and is the model, for instance, for Yumitani’s (1998) description of Jemez; Watkins also has a number of later publications both on Kiowa (Watkins, 1990, 1993) and on the family (Watkins, 1996). Good access points to the rich literature on Southern Tiwa agreement and incorporation are Allen et al. (1990) and Rosen (1990), though one should not neglect subsequent works by Frantz (e.g., 1991, 1995, 2009), which provide further fine detail. For Tewa, the best description remains Speirs’ (1966) dissertation.
For more recent theoretical perspectives on Kiowa-Tanoan linguistics, one can read Adger et al. (2009) on Kiowa clause structure, Heck and Richards (2010) on Southern Tiwa agreement restrictions, and Harbour (2011b), Watanabe (2015) on Jemez and Kiowa noun classes.
Adger, D., Harbour, D., & Watkins, L. (2009). Mirrors and microparameters: Phrase structure beyond free word order. Cambridge, U.K.: Cambridge University Press.Find this resource:
Allen, B., Frantz, D., Gardiner, D., & Perlmutter, D. (1990). Verb agreement, possessor ascension, and multistratal representation in Southern Tiwa. In B. Joseph & P. Postal (Eds.), Studies in relational grammar 3 (pp. 321–383). Chicago: University of Chicago Press.Find this resource:
Allen, B., Gardiner, D., & Frantz, D. (1984). Noun incorporation in Southern Tiwa. International Journal of American Linguistics, 50, 292–311.Find this resource:
Baker, M. (1996). The polysynthesis parameter. Oxford: Oxford University Press.Find this resource:
Frantz, D. (1991). Null heads and noun incorporation in Southern Tiwa. In D. J. Costa (Ed.), Proceedings of the 22nd Annual Meeting of the Berkeley Linguistics Society: Papers from the Special Session on Historical Issues in Native American languages (pp. 32–38). Berkeley, CA: Berkeley Linguistics Society.Find this resource:
Frantz, D. (1995). Southern Tiwa argument structure. In C. S. Burgess, K. Dzi-wirek, & D. Gerdts (Eds.), Grammatical relations: Theoretical approaches to empirical questions (pp. 73–95). Stanford, CA: CSLI Publications.Find this resource:
Frantz, D. (2009). Constraints on verb stacking in Southern Tiwa. In D. Gerdts, J. Moore, & M. Polinsky (Eds.), Hypothesis A/hypothesis B: Linguistic explorations in honor of David M. Perlmutter (pp. 173–181). Cambridge, MA: MIT Press.Find this resource:
Harbour, D. (2007). Morphosemantic number: From Kiowa noun classes to UG number features. Dordrecht, Netherlands: Springer.Find this resource:
Harbour, D. (2011a). Kiowa-Tanoan languages. Oxford Bibliographies Online.
Harbour, D. (2011b). Valence and atomic number. Linguistic Inquiry, 42, 561–594.Find this resource:
Harbour, D. (2013). “Not plus” isn’t “not there”: Bivalence in person, number, and gender. In O. Matushansky & A. Marantz (Eds.), Distributed Morphology today (pp. 135–150; references, 223–249). Cambridge, MA: MIT Press.Find this resource:
Harbour, D., Rinaldi, M., Skene, M., & Tinay, L. (2014). Tiwa tales: From the work of Carobeth Tucker Harrington. Queen Mary’s OPAL 31.
Harrington, J. P. (1916). Ambiguity in the Taos personal pronoun. In W. H. Holmes, E. Leary, & F. W. Hodge (Eds.), Holmes anniversary volume: Anthropological essays presented to William Henry Holmes in honor of his seventieth birthday, December 1, 1916 (pp. 142–156). Washington, DC: J. W. Bryan.Find this resource:
Harrington, J. P., & Roberts, H. H. (1928). Picurís children’s stories with texts and songs. In J. W. Fewkes (Ed.), Forty-third annual report of the Bureau of American Ethnology to the Secretary of the Smithsonian Institute: 1925–1926 (pp. 289–447). Washington, DC: United States Government Printing Office.Find this resource:
Heck, F., & Richards, M. (2010). A probe-goal approach to agreement and incorporation restrictions in Southern Tiwa. Natural Language and Linguistic Theory, 28, 681–721.Find this resource:
Noyer, R. (1992). Features, positions, and affixes in autonomous morphological structure (Doctoral dissertation) Massachusetts Institute of Technology.Find this resource:
Rosen, C. (1990). Rethinking Southern Tiwa: The geometry of a triple agreement language. Language, 66, 669–713.Find this resource:
San Juan Pueblo Bilingual Program. (1982). Téwa pehtsiye: Tewa tales, San Juan dialect. Portales, NM: Bishop Publishing.Find this resource:
Speirs, R. H. (1966). Some aspects of the structure of Rio Grande Tewa (Unpublished doctoral dissertation). State University of New York, Buffalo.Find this resource:
Sprott, R. W. (1992). Jemez syntax (Unpublished doctoral dissertation). University of Chicago.Find this resource:
Sutton, L. (2010). Noun class and number in Kiowa-Tanoan: Comparative-historical research and respecting speakers’ rights in fieldwork. In A. L. Berez, J. Mulder, & D. Rosenblum (Eds.), Fieldwork and linguistic analysis in indigenous languages of the Americas (pp. 57–89). Honolulu, HA: University of Hawai‘i Press.Find this resource:
Sutton, L. (2014). Kiowa-Tanoan: A synchronic and diachronic study (Unpublished doctoral dissertation). University of New Mexico.Find this resource:
Tucker Harrington, C. (1920a). Isleta language: Texts and analytical vocabulary. Washington, DC: Bureau of American Ethnology.Find this resource:
Tucker Harrington, C. (1920b). The Isleta pronoun. Washington, DC: Bureau of American Ethnology.Find this resource:
Watanabe, A. (2015). Valuation as deletion: Inverse in Jemez and Kiowa. Natural Language and Linguistic Theory, 33(4), 1387–1420.Find this resource:
Watkins, L. J. (1984). A grammar of Kiowa. Lincoln: University of Nebraska Press. With the assistance of Parker McKenzie.Find this resource:
Watkins, L. J. (1990). Noun phrase versus zero in Kiowa discourse. International Journal of American Linguistics, 56, 410–426.Find this resource:
Watkins, L. J. (1993). The discourse functions of Kiowa switch-reference. International Journal of American Linguistics, 59, 137–164.Find this resource:
Watkins, L. J. (1996). Reconstructing person and voice in Kiowa-Tanoan: Pitfalls and progress. In D. Librik & R. Beeler (Eds.), Proceedings of the 22nd Annual Meeting of the Berkeley Linguistics Society: Papers from the Special Session on Historical Issues in Native American languages (pp. 139–152). Berkeley, CA: Berkeley Linguistics Society.Find this resource:
Yumitani, Y. (1998). A phonology and morphology of Jemez Towa (Unpublished doctoral dissertation). University of Kansas.Find this resource: