One of the most interesting things to me highly morphologically complex languages, such as, for example, Plains Cree, is the sheer volume of so-called ‘minor phenomena’; small, often highly context specific morphosyntactic processes which are often overlooked in linguists’ frenzied attempts to catalogue the broader strokes of a language’s grammar. One particular ‘minor phenomenon’ in Plains Cree that I have been interested in for some time is the so-called Inanimate Actor paradigm, a set of compound suffixes used to encode inanimate subjects in transitive verbs which has, to my knowledge, been only very tersely described in existed documentary literature on the language. I wrote a paper describing the usage of this paradigm in corpora last year for a morphology course, but for a variety of reasons never ended up publishing the results. Although I still plan on eventually publishing this material, in the interim, you can read a truncated version of the paper here:
A DESCRIPTIVE OVERVIEW AND PROVISIONAL CxM SCHEMATISATION OF THE INANIMATE ACTOR SUBPARADIGM IN PLAINS CREE (nêhiyawêwin)
ABSTRACT
Plains Cree (ISO:crk), an inflectionally rich, polysynthetic Algonquian language spoken primarily in Saskatchewan and Alberta, segments its lexicon of verbs into four inflectional subclasses based on the transitivity of the verb and the animacy of its arguments. These subclasses are VII (Inanimate Subject Intransitive), VAI (Animate Subject Intransitive), VTI (Animate Subject-Inanimate Object Transitive), and VTA (Animate Subject-Animate Object Transitive). However, conspicuously absent from these permutations are transitive verbs with inanimate subjects. Rather than creating a fifth subclass to describe such events, Cree speakers instead use an inflectional subparadigm of VTAs known as the Inanimate Actor (or IA). This investigation seeks to provide both a descriptive and theoretical account of the IA subparadigm, recording the various structural and semantic characteristics associated with its use in corpora and providing a provisional schematisation of the construction in the framework of Construction Morphology (Booij 2010).
1. INTRODUCTION
1.1 Plains Cree
Plains Cree (ISO:crk, known endonymically as nêhiyawêwin) is an Algonquian language spoken across Alberta, Saskatchewan, the Northwest Territories, and Montana. With ~30 000 native speakers (Statistics Canada 2017), it is the most widely spoken language in the Cree dialect continuum, and one of the most widely spoken Indigenous languages in Canada, both geographically and demographically. Typologically, Plains Cree is highly polysynthetic, with a particular richness of verbal inflection and derivation. Nominal morphology is generally simpler, with nouns being governed by a binary system of grammatical gender based on animacy:
(1) ninôhtê-nakiskawâw
ni-nôhtê=nakiskaw-âw
1sɢ>3sɢ.ɪɴᴅ-want=greet(ᴠᴛᴀ)-1sɢ>3sɢ.ɪɴᴅ
‘I want to greet him’
Cree verbs may be divided into four inflectional subclasses depending on their transitivity and the animacy of their arguments. These subclasses are VII (Inanimate Subject Intransitive), VAI (Animate Subject Intransitive), VTI (Animate Subject-Inanimate Object Transitive), and VTA (Animate Subject-Animate Object Transitive). All four subclasses display affixal person-number agreement with their subjects, with VTAs also agreeing with the person and number of their objects. These agreement affixes may be drawn from one of two distinct, separate sets; the Independent and the Conjunct. Verbs use agreement affixes from the Independent set when they head a matrix clause and Conjunct affixes when they head a subordinate clause. However, verbs may also take Conjunct affixation in matrix clauses if the information which they convey is not verifiable; in this way, Conjunct affixes may also serve as dubitative evidential markers:
(2) Jeff kinosiw
Jeff kinosi-w
Jeff be.tall(ᴠᴀɪ)-3sɢ.ɪɴᴅ
‘Jeff is tall’ (witnessed by speaker, firsthand information) (Cook 2007)
(3) nikiskêyihtên Jeff ê-kinosit
ni-kiskêyiht-ên Jeff ê-kinosi-t
1sɢ.ɪɴᴅ-know(ᴠᴛɪ)-1sɢ.ɪɴᴅ Jeff ᴄɴᴊ-be.tall(ᴠᴀɪ)-3sɢ.ᴄɴᴊ
‘I know (that) Jeff is tall’
(4) Jeff ê-kinosit
Jeff ê-kinosi-t
Jeff ᴄɴᴊ-be.tall(ᴠᴀɪ)-3sɢ.ᴄɴᴊ
‘Jeff is tall’ (hearsay, secondhand information) (Cook 2007)
A conspicuous syntactic and semantic gap among the four previously mentioned verbal inflectional subclasses is that of transitive verbs with inanimate subjects. This gap in animacy-transitivity permutations is not entirely unprecedented; events wherein inanimate objects act on animate ones are relatively uncommon cross-linguistically, and tend to be seen as marked constructions (Yamamoto 1999). However, despite lacking a dedicated inflectional subclass for such events, Plains Cree can still encode inanimate subjects on transitive verbs using a subparadigm of VTAs; namely, the Inanimate Actor.
1.2. The inanimate actor
The Inanimate Actor subparadigm (henceforth IA) is a set of inflectional affix combinations modifying VTAs so as to enable them to encode for inanimate subjects. This is accomplished by adding the suffix –iko to the relevant VTA stem and then inflecting the resultant base with person-number affixes from the VAI subclass to indicate the person and number of the object (see Tables A1 and A2 in the Appendix):
(5) mêtoni ê-mamihcihit
mêtoni ê-mamihcih-it
really ᴄɴᴊ-make.proud(ᴠᴛᴀ)-3sɢ>1sɢ.ᴄɴᴊ
‘s/he really makes me proud’
(6) wahwâ, iyikohk ê-mamihcihikoyân ôma niy âyi
wahwâ, iyikohk ê-mamihcih-iko-yân ôma niy âyi
oh.my so.much ᴄɴᴊ-make.proud(ᴠᴛᴀ)-ɪɴᴀɴ-1sɢ.ᴄɴᴊ ᴅᴇᴍ(ɴɪ) 1sɢ thing
‘Oh my, that there really makes me proud’ (context; discussing a trophy)
(Bear et al. 1998)
Despite filling an otherwise vacant syntactic niche in the language, the IA subparadigm has only been tersely recorded in extant documentary literature on Cree. It is broadly overlooked in pedagogical texts (Okimāsis 2004), and the two major academic sources which discuss it, Wolfart (1973) and Wolvengrey (2011b), both provide only brief summaries and (mutually contradictory) paradigm tables. As such, the following investigation intends to serve both as an exploration of potential theoretical explanations for the IA and as a thorough, usage-based descriptive account of the distribution, semantic prosody, and idiosyncrasies of the construction.
2. METHODOLOGY
Examples of the IA were extracted from a set of three morphologically-tagged Plains Cree corpora, totalling 174 340 tokens. The first of these corpora was the Ahenakew-Wolfart Corpus (AW), a geographically diverse set of monolingual personal reminiscences collected throughout the 1980s and 1990s, totalling 81 565 tokens (17 976 types). The second was the Bloomfield Corpus (BF), a collection of rehearsed and spontaneous stories (and their English translations) collected from monolingual Cree speakers at Sweetgrass First Nation, Saskatchewan during the 1920s and 1930s; this corpus totalled 72 475 tokens (15 267 types). The final source was the Maskwacîs Speech Database (SD), a bilingual collection of elicited words and phrases compiled in Maskwacîs, a Cree community in Central Alberta, between 2014 and 2018, totalling 20 300 entries. Out of these, only AW is written entirely in the modern Standard Roman Orthography for Plains Cree; BF is written in an outdated variant of the SRO, and, as of the time of writing, only 67.1% of entries in SD are orthographically standardised.
Instances of the IA were located through searching the relevant morphological tags; additionally, the source documents for each corpus were searched using regular expressions for all possible IA-person-number morpheme combinations. Finally, unclear or ambiguous instances of the IA were parsed by a native speaker for verification.
- FINDINGS
3.1 Quantitative Overview
Searching using only corpus-internal morphological tagging yielded 101 examples of the IA across all three corpora (28 in AW, 21 in BF, 51 in SD). However, when this search is expanded with regular expressions, a total of 164 can be found (66 in AW, 38 in BF, 60 in SD). Among these, a total of 91 unique verbal lexemes are used as stems, the majority of which being hapax legomena. Further, a total of 70% of IA verbs in the corpora used Conjunct person-number affixes, compared with only 51% of VTAs as a whole in corpora taking Conjunct affixation (66% in AW and 38% in BF, SD being excluded from these counts due to it being a compilation of isolated, elicited words and sentences, rather than running text).
3.2 Collocational Preferences
The IA shows clear collocational preferences for a small set of VTA stems, although only one verb stem (sêkih-) occurs with IA agreement in all three corpora:
-
- nipah– (‘to kill s.o.’): 18 (15 in BF, 3 in AW, 0 in SD)
- mamitonêyih– (‘to make s.o. think’): 6 (0 in BF, 4 in AW, 2 in SD)
- sêkih– (‘to frighten s.o.’): 5 (3 in BF, 2 in AW, 1 in SD)
- astâh– (‘to frighten s.o.’): 3 (2 in BF, 0 in AW, 1 in SD)
- misiwanâcih– (‘to ruin s.o.’): 3 (1 in BF, 2 in AW, 0 in SD)
- wawânêyihtamih– (‘to make s.o. desperate’): 3 (2 in AW, 1 in SD)
- kiskisoh– (‘to remind s.o.’): 6 (0 in BF, 0 in AW, 6 in SD)
IA verbs also show a definite preference for pronominal subjects, with 53.8% taking an overtly realised inanimate pronoun as a subject:
- kîkway (‘something’): 27.9%
- êwako (‘this’): 9.61%
- ôma (‘this’): 7.69%
- nama kîkway (‘nothing’): 4.8%
- anima (‘that’): 3.84%
By comparison, IA verbs with no overtly marked subject are uncommon, with only 20 instances (19.2%, again excluding SD) across the corpora. This is of particular note due to the overwhelming tendency of Cree verbs towards subject elision; 70.3% of all VTAs across the corpora lack an overtly marked subject (Schmirler forthcoming). This scarcity of subjectless IA verbs is most likely the product of constructional markedness; being a marked construction requiring a specific, comparatively uncommon context to licence its use (as opposed to the use of the more common ‘standard’ VTA paradigm), speakers tend towards providing a specific, contextual explanation for the IA in the form of an overt subject.
In addition to their collocational preferences in verb stems and subjects, IA verbs also have distinct preferences in their distribution with other morphemes. Namely, verbs with IA marking have a clear tendency towards taking objects with singular number. Even more strikingly, IA verbs with obviative objects are entirely unattested in corpora, despite obviative objects being the most common objects for VTAs as a whole. This directly contradicts the claims of one of the two principal documentary sources on the IA (Wolvengrey 2011b), which lists the obviative as a possible object for IA verbs.
| IA Verbs in Corpora | All VTAs in Corpora | ||||
|---|---|---|---|---|---|
| Object | Frequency | Proportion | Frequency | Proportion | |
| 3sɢ | 44 | 26.0% | 2237 | 26.6% | |
| 3ᴘʟ | 22 | 13.0% | 382 | 4.56% | |
| 2sɢ | 49 | 29.0% | 517 | 6.1% | |
|
2ᴘʟ |
2 | 1.18% | 86 | 1.02% | |
|
1sɢ |
42 | 24.8% | 528 |
6.30% |
|
|
1ᴘʟ |
5 | 2.95% | 45 |
0.0536% |
|
|
12ᴘʟ |
4 | 2.36% | 46 |
0.0548% |
|
|
4 |
0 | 0% | 4521 |
33.6% |
|
| X | 1 | 0.592% | 20 |
0.0239% |
|
Table 1: Distribution of object person-number for IA verbs and VTAs in general. For example, there 49 total instances of IA verbs with a 2sɢ object in corpora (constituting 29% of total IA verbs in corpora), compared with 517 instances of a 2sɢ object on all VTAs (constituting 6.1% of all total VTAs in corpora)
3.3 Semantics
Verbs taking IA affixation have a general tendency to describe events exhibiting negative semantic prosody, often being used to refer to actions causing pain, distress, or destruction. Of the 98 IA verbs with English translations in the corpora (that is, those from BF and SD), a total of 52 carried negative prosodic meanings, compared with 25 neutral and 22 positive.
(7) nikwatakihikon nistikwân
ni-kwatakih-iko-n ni-stikwân
ɪɴᴀɴ>1sɢ.ɪɴᴅ-torment-ɪɴᴀɴ-1sɢ.ɪɴᴅ 1sɢ.ᴘᴏss-head(ɴɪ)
‘My head greatly torments me’ (Bloomfield 1934)
(8) tâpiskôc ôma minihkwêwin kâ-misiwanâcihikocik
tâpiskôc ôma minihkwêwin kâ-misiwanâcih-iko-cik
like ᴅᴇᴍ(ɴɪ) drinking(ɴɪ) ᴄɴᴊ-ruin(ᴠᴛᴀ)-ɪɴᴀɴ-3ᴘʟ.ᴄɴᴊ
‘Just as this drinking takes them to ruin’ (Bear et al. 1998)
3.4 Speaker Perception
Being a marked construction, speakers find IA verbs without supporting context infelicitous and difficult to parse. However, when placed in context, the same speakers have no difficulty understanding and accepting IA forms as grammatical. A transcribed excerpt from an elicitation session, provided in the Appendix as A2, demonstrates a typical speaker reaction to an IA verb out of context.
- DISCUSSION AND CONCLUSIONS
Despite its near non-existence in extant documentary literature of Plains Cree, corpus analysis reveals the IA subparadigm to be both productive (owing to the large number of hapax legomena verb stems with which it occurs) and parseable by speakers, with novel generalisations on the schema being readily possible. Although relatively infrequent (owing, if nothing else, to the pragmatic infrequence of inanimate transitive subjects), the lack of documentation on the IA thus represents a major gap in academic and pedagogical literature, in turn impacting the quality of language resources and tools which linguists are able to create based on existing documentation.
- REFERENCES
Ahenakew, F. (1987). wāskahikaniwiyiniw-ācimowina / Stories of the House People. University of Manitoba Press
Bear, G., Fraser, M., Calliou, I., Wells, M., Lafond, A., Longneck, R. (1998). Kōhkominawak otācimowiniwāwa / Our grandmothers’ lives as told in their own words. University of Regina Press.
Bloomfield, L. (1934). Plains Cree texts. G. E. Stechert & Co.
Booij, G. (2010). Construction morphology. Language and Linguistics Compass. 4(7):543-555. doi:10.1111/j.1749-818X.2010.00213.x.
Cook, C. (2007). Distinguishing modes in Plains Cree. In H. C. Wolfart (Ed.), Papers of the Thirty Eighth Algonquian Conference (pp. 47-80). University of Manitoba Press.
Cook, C. (2008). The syntax and semantics of clause-typing in Plains Cree. [Doctoral thesis, University of British Columbia]. University of British Columbia Library. https://open.library.ubc.ca/soa/cIRcle/collections/ubctheses/24/items/1.0066456
Masuskapoe, C. (2010). piko kīkway ē-nakacihtāt: kēkēk otācimowina ē-nēhiyawastēki mitoni ē-āh-itwēt māna Cecila Masuskapoe / There’s nothing she can’t do : Kēkēk’s autobiography published in Cree. Exactly as told by Cecilia Masuskapoe. In H. C. Wolfart & F. Ahenakew (Eds.), Algonquian and Iroquoian linguistics memoirs. University of Manitoba Press.
Okimāsis, J. (2004). Cree: Language of the plains. University of Regina Press.
Schmirler, K. (forthcoming). Syntactic features and text types in 20th century Plains Cree: A Constraint Grammar approach. [Doctoral thesis, University of Alberta].
Statistics Canada. (2017). Aboriginal languages in Canada, 2016 Census of Population.
www150.statcan.gc.ca/n1/pub/11-627-m/11- 627-m2017035- eng.htm
Wolfart, H. C. (1973). Plains Cree: a grammatical study. Transactions of the American Philosophical Society, 63(5):1-90. doi:10.2307/1006246
Wolvengrey, A. (2011a). Cree: Words. University of Regina Press.
Wolvengrey, A. (2011b). Semantic and pragmatic functions in Plains Cree syntax [Doctoral thesis, University of Amsterdam]. UvA-DARE (Digital Academic Repository). https://dare.uva.nl/search?identifier=153f3405-f029-4046-8d95-9f2a087b4123
Yamamoto, M. (1999). Animacy and reference: A cognitive approach to corpus linguistics. John Benjamin’s Publishing Company.
ACKNOWLEDGEMENTS
I would like to thank the Social Sciences and Humanities Research Council of Canada (SSHRC), who have funded and provided ethics approval for this study through their Connections Outreach Grant (611-2016-0207) and Partnership Development Grant (890-2013-0047) for 21st Century Tools for Indigenous Languages (21CT4IL).
APPENDIX
A1. Glossary of Abbreviations not in Common Use
1ᴘʟ – Exclusive 1ᴘʟ (‘we, not including you’)
12ᴘʟ – Inclusive 1ᴘʟ (‘we, including you’)
4 – Obviative third person
ᴀɢʀ – Agreement (for person and number)
ᴄɴᴊ – Conjunct mode
ɪɴᴀɴ – Inanimate actor
ɪɴᴅ – Independent mode
ɪᴛᴇʀ – Iterative
ɴᴀ – Animate noun
ɴɪ – Inanimate noun
ᴏʙᴠ – Obviative noun
ᴠɪɪ – Inanimate subject intransitive verb
ᴠᴀɪ – Animate subject intransitive verb
ᴠᴛɪ – Animate subject, inanimate object transitive verb
ᴠᴛᴀ – Animate subject, animate object transitive verb
X – Unspecified actor
A2. Excerpt from an elicitation session between the author (D) and a native-speaker consultant (R) conducted 2/12/22, discussing the following sentence:
(12) mihko ê-nahêyihtamihikot
mihko ê-nahêyihtamih-iko-t
blood(ɴɪ) ᴄɴᴊ-satisfy(ᴠᴛᴀ)-ɪɴᴀɴ-3sɢ.ᴄɴᴊ
s/he is bloodthirsty (lit. ‘blood satisfies him/her’)
- R) Well this just doesn’t make sense … ‘he is bloodthirsty’, ‘blood makes him feel good’, and… yeah?
- D) You said this one doesn’t make sense?
- R) Well, why would somebody… well, I suppose..
- D) If you were talking about a vampire or something, or perhaps a wolf or something like that, they want to eat, you might say mihko ê-nahêyihtamihikot?
- R) Yeah, it makes him feel good, ê-nahêyihtamihikot (yawn)
- D) So does this sentence make sense from a grammatical point of view? I mean, maybe it’s a weird thing to say about someone, but-
- R) It’s a weird thing to say, but grammatically it’s fine, mihko ê-nahêyihtamihikot
- D) Would there be another way that you would know of saying this, how you would say-
- R) [laughs] I honestly don’t know, this is the first I’ve heard of this… I wish I knew the context where, what they were talking about
- D) If I said this instead, if I said it like mihko ê-nahêyihtamihât*, would that make any sense-
- R) môya [no]
- D) That doesn’t make any sense?
- R) mihko ê-nahêyihtamihikot is, uhh…
[tangent]
- D) So, ‘blood satisfies him’?
- R) ‘Blood satisfies him’, then we’re talking about, yeah, we could be talking about a wolf, we could be talking about a vampire, we could be talking about, [yawn], um, something
*ê-nahêyihtamihât would be the expected form for a grammatically animate subject
| Independent | Conjunct | |||||
| Preverb | Stem | Suffix | Preverb | Stem | Suffix | |
| 1sɢ | ni- | -ikon | ê- | -ikoyân | ||
| 2sɢ | ki- | -ikon | ê- | -ikoyan | ||
| 3sɢ | -ikow | ê- | -ikot | |||
| 1ᴘʟ | ni- | -ikonân | ê- | -ikoyâhk | ||
| 12ᴘʟ | ki- | -ikonânaw | ê- | -ikoyahk | ||
| 2ᴘʟ | ki- | -ikonâwâw | ê- | -ikoyêk | ||
| 3ᴘʟ | -ikowak | ê- | -ikocik | |||
| 4 | -ikoyiwa | ê- | -ikoyit | |||
Table A1: The IA subparadigm for VTAs (adapted from Wolvengrey 2011b)
| Independent | Conjunct | |||||
| Preverb | Stem | Suffix | Preverb | Stem | Suffix | |
| 1sɢ | ni- | -n | ê- | -yân | ||
| 2sɢ | ki- | -n | ê- | -yan | ||
| 3sɢ | -w | ê- | -t | |||
| 1ᴘʟ | ni- | -nân | ê- | -yâhk | ||
| 12ᴘʟ | ki- | -nânaw | ê- | -yahk | ||
| 2ᴘʟ | ki- | -nâwâw | ê- | -yêk | ||
| 3ᴘʟ | -wak | ê- | -cik | |||
| 4 | -yiwa | ê- | -yit | |||
Table A2: The basic paradigm for VAIs (Okimāsis 2004)