First posted: 23 Nov 2014
Last revised: 24 Mar 2018 (minor copy edits: 26 Dec 2020)
URL: http://ezlinguistics.blogspot.de/p/kalapuya-verb-prefixes.html
1. Introduction
The Kalapuyan languages are a small family of Native American languages that were spoken in the Willamette Valley (Oregon). Internally, the family can be divided into three branches: Northern, Central and Southern Kalapuyan. The Kalapuyan family is usually included as part of the proposed Penutian macrofamily. The closest relative of the Kalapuyan languages seems to be Takelma, based on evidence from shared lexicon and bound morphemes, although this remains controversial.
The main source for Kalapuyan languages is the large corpus of texts published by Melville Jacobs (1945). The bulk of these texts is from Central Kalapuyan: Santiam texts were collected by Jacobs himself, whereas material from the P̓īnefu (= Mary's River) and Lower McKenzie dialects were originally collected by Leo Frachtenberg in the 1910s. A first analysis of the grammar of Santiam Kalapuya was undertaken by Jonathan Banks (2007), concentrating on the verbal morphology of Santiam. This paper is based on Banks' article, and also on my own comprehensive analysis of Jacobs' text corpus. In several points, I arrive at conclusions different from Banks' results.
In Kalapuyan languages, every verb is expanded to the left by diverse inflectional prefixes, indicating tense, subordination, person/number and location/direction. Completely lacking are prefixes with "concrete" meaning (i.e. expressing body-parts, shape etc.) which are quite common in other languages of the area, including Takelma.
The concatenation of these prefixes is quite transparent and regular, with only few phonetic variations due to regular morphophonological processes. This is quite in contrast with the rather complex and "synthetic" nature of verb suffixation at the opposite side. We can therefore assume that the leftward verb affixation developed at a relatively late stage, probably through capture (via clitization) of earlier independent auxilliaries. Contact with Athabascan and Chinookan languages may have played a role in this development. Note that the presumed closest relative of Kalaypuan, i.e. Takelma, has almost no inflectional prefixes; rather, all verbal inflection for person and TAM involves highly "synthetic" suffixes.
1.1 Notes on spelling and phonology
In the text corpus collected by Jacobs, a relatively narrow transcription is employed. The most prominent example is the representation of phonemic /a/ as a or e (spelled "ɛ" by Jacobs) in the Santiam texts. I will follow Jacobs's spelling when citing examples with a or e, but in the formal description, I will always use phonemic a.
For the obstruents, I follow Jacobs' spelling of unaspirated lenes as b/d/g, but without small caps. Aspirated fortis stops are indicated by p/t/k. Affricates are respelled as c/c̓.
In many unstressed syllables, including prefixes, the contrast between short i and u appears to be non-phonemic. This can be exemplified by the past tense prefix gi-/gu-. The back variant u appears preceding a labial consonant, whereas i appears preceding alveolar consonants:
The prefix gum- changes to gin- when final m assimilates to an alveolar consonant in a following prefix:
The front variant i often (especially in the P̓īnefu dialect) becomes zero in open syllables, e.g.
The same pattern is also found with the first person plural future tense prefix dum-, the SAP present tense prefix cu-/ci-, and the third person present tense prefix um-. In the prefix cu-/ci-, the vowel is regularly dropped preceding the alveolar stop d, resulting in sda-, sdū-.
Neutralization of short high vowels is also found with the shortened variants of the first and second person plural subject prefixes dū- and dīp-: for the former, we have short du- and di-, with the latter variant dominating; dīp- is always shortened to dup-, having back u triggered by the final labial stop.
Since the variation between i and u is largely conditioned by the phonetic environment, I propose a single high central vowel phoneme here, viz. /ɨ/. The above-listed prefixes will thus be spelled:
No variation is found with the third person plural prefix ni-, which invariably has a front vowel.
2. Banks' analysis of Santiam verb prefixes
Banks proposes 11 prefix positions, which can be tabulated as follows:
I mainly differ in the analysis of the prefixes appearing in positions 1 to 7. The distinction of seven positions preceding the translocative prefix d(i)‑ is over-atomistic, and some of the formative elements are in part incorrectly identified. Basically, the elements filling Banksʼ positions 1 to 3 indicate tense, partially distinguishing person, while the elements of positions 4 to 7 seven indicate syntactic function. The following section 3 will elaborate on this in detail.
The prefixes appearing in Banks' position 8 to 11 will be discussed in section 4. In Section 5, I will describe the copula prefix ihi-/i-, which primarily appears with unpossessed nouns in equative clauses. This prefix has not been described to my knowlegde so far in earlier research.
3. Macro-Position 1: tense-marking and clause relational prefixes (Bank's positions 1 to 7)
In place of Banks' positions 1 to 7, I posit one macro-position 1, which contains both mono- and poly-morphemic prefixes. Most prefixes in macro-position 1 indicate tense (out of six tenses) and clause-related function (independent clause or one out of four dependent clause types), and can be arranged into two "neat" major paradigms. Two tense-neutral prefixes appear in embedded clauses functioning as same-subject and switch-subject infinitive.
3.1. Tense-marking prefixes
Santiam Kalapuya distinguishes six tenses. Their function can be abstracted from the text corpus.
The first four tenses are expressed by person-neutral prefixes, while in present tense, third person forms contrast with common forms for first and second person (speech act participants = SAP). In future tense, there is a distinct prefix for all three persons.
The tense-relational-markers of macro-position 1 are composed of one or two prefixes and can be divided into two classes of paradigms, based on their internal structure:
Class 1: non-future tenses
Class 2: future
3.1.1. Non-future tenses
Tense-relational-markers of Class 1 can be tabulated as follows:
The non-relative forms can be analyzed in a straightforward manner:
We can identify the following formative elememts:
Prefix components in Position A indicate tense, composed of a characteristic consonant followed by a thematic vowel, while the elements in Position B indicate the function of the clause as independent or one out of three dependent clause types. As we can see, the habitual and recent past prefixes carry the same thematic vowel a, whereas the past and present prefixes agree in taking the vowel ɨ.
The prefix -da- appears as variant of locative -dū-, exclusively in the position preceding the spatial prefix han-: in the Santiam corpus, we find several instances of -de-hen- next to "regular" -dū-hen-; in the P̓īnefu texts, we regularly find contracted -dan- corresponding to Santiam -dū-hen- or -de-hen-.
Note that third person present is zero-marked with the locative prefix -dū-. The independent prefix ɨm- also has a zero-variant when it precedes a prefix element beginning with a nasal, viz. ni- "third person plural" and ma- "cislocative", e.g. ma-lami "he brings it inside", ni-dahai "they kill him".
The relative prefixes are used in relative and cleft clauses relating to subject and object (e.g. gan-hūkni "[the one] who ate it" or "[the thing] that he ate"). Like the basic tense prefixes, they are composed of the actual tense-marking consonant prefix, and a characteritic thematic element.
We thus find the same pattern as with the basic tense prefixes: habitual and recent past agree in taking -ī- as formative element, next to past and SAP present which have -an-.
The prefixes can- and gan- are analyzed as c-a-m- and g-a-m- by Banks, contaning the "assertive" preifx a- plus the "finite" prefix m-. The final nasal is however definitely n, as can be seen from examples such as gan-hūkni. The vocalic part a- is analyzed as the same element as in the recent past prefix gʷa-. Since these two environments are actually diametrically opposite, I take -an- as an undividable relative marking component.
3.1.2. Future tense prefixes
Class 1c contains future tense markers distiguishing first, second and third person. Here, we find some signifacant differences in the paradigms of Santiam and P̓īnefu.
The first person prefixes largely agree, except for the loss of initial d in the subjunctive form in P̓īnefu. Loss of d is also found preceding the first person plural prefix dū-, resulting in indū-.
Second person prefixes have different characteristic consonants: Santiam has n, whereas P̓īnefu has the stop d. External evidence from Northern Kalapuyan (da-) points at *d for Proto Central Kalapuyan.
The most significant difference is found with the subjunctive prefixes for second and third person. In P̓īnefu, we find the same pattern as in the remaining tense forms, where the final nasal m found in the independent prefixes is replaced by the subjunctive marker da-.
past
2 future
3 future
The Santiam Kalapuya forms however are characterized by having final m as an inseparable part of the future tense prefixes:
past
2 future
3 future
The formative elemenents of the future prefixes can thus be tabulated as follows:
Position A
1 future
2 future
3 future
3.2. Tense-neutral subordinating prefixes
Next to the tense marking prefixes, Santiam has two tense-neutral subordinating prefixes. The infinitive prefix dumī‑ marks same subject complement clauses, e.g.
In Santiam, the infinitive prefix is homophonous to the first preson future conditional. In P̓īnefu, the infinitve prefix has a shortened variant dī‑, which occurs much more often than dumī‑; the first preson future conditional prefix however is unvariably dumī‑.
The subjunctive prefix gī- has two functions. First, it is used in hypothetical conditional clauses, further it appears in different-subject complement clauses:
3.3. Imperative
The imperative is marked by default by the subjunctive prefix da- without any preceding tense marker.
It can cooccur with the translocative prefix di-/t- (position 2A) and the third person plural prefix ni- (position 2B).
With the second person plural prefix dɨp- (position 2B) and the cislocative prefix ma- (position 2C), the imperative is zero marked.
4. Macro-position 2 (Banks' position 8 to 11)
Banks posits eight prefixes filling his positions 8 to 11: t-/d-/di-, de-/den-, du-/di-, dup-/dip-, ni-, ma-/man-, he-/hen- and ci-/ce-/cu-. I largely agree with Banks' analysis, except for three points:
- The nasal is an integral part of the prefixes hen- and den-. It assimilates to a following stop and is dropped preceding f and the sonorants m, n, l, w and y.
- The prefix den- is not a basic prefix, but a portmonteau form of di- and hen- (dɨ- and han- in my phonemic spelling).
- There are two distinct prefixes ma- and man-. The prefix ma- is invariant in all positions and also appears before stops, whereas man- has an homophonous allomorph ma- that is used in the same environment as the allomorphs he- and de-. Further, based on its distribution, man- can be analyzed as a portmonteau of the indicative prefix -m- (position 1B) and han-.
All combinations of these prefixes that are found in the text corpus can be arranged in the following grid:
We can see that dɨ-, ma- and cɨ- are mutually exclusive, even though they appear in different positions. Likewise, han- and the plural person markers cannot cooccur. The combination dīp-cɨ- is unattested but this must be due to the relataively low text frequency of cɨ- in the corpus as compared to dɨ- and ma- (cf. dū-cɨ-, which only occurs once in the whole corpus). The only real gap in the grid (there is no han- + ma-) is the result of a funcional constraint against ma- appearing with any spatial-related affix, i.e. not just with han- as seen in the grid but also with the locative clause-marking prefix dū- (position 1B).
Based on this grid, I posit three sub-positions for macro-position 2: 2A and 2C for the directional prefixes dɨ-, ma- and cɨ-, and 2B for the person marking prefixes and han-.
4.1. Position 2A: The translocative prefix dɨ-
The "translocative" prefix indicates movement away from the deictic center (speaker, or the main participant in a narrative) towards a further reference point, similar to German hin.
In most cases, the prefix appears as di- in the corpus. The apocopated variant also occurs frequently, whereas the variant du- is only found preceding the second person plural marker dɨp-.
4.2. Position 2B: The plural person markers
The first and second person plural prefixes dū- and dīp- always refer to the subject of the clause, whereas the third person plural prefix nī- can refer to subject or object. Dū-, dīp- and nī- each have a shortened variant, viz. dɨ-, dɨp-, which are subject to neutralization of short high vowels in unstressed position, and ni- which invariably has the vowel i. The second person plural prefix further drops the final stop when preceding the cislocative prefix ma-, resulting in dɨ-ma-.
In present and future tense, dū- and dīp- are always bound to the corresponding tense-relational prefix:
The prefix nī- however can coocur with first and second person tense-prefixes, in which case it expresses a third person plural object:
4.3. Position 2B: The locative marker (h)an-
The locative marker (h)an- also appears in position 2B. In Banks' analysis, the "proximal" prefix han- occupies the position immediately following the plural person markers, together with cislocative ma-. A synopsis of all occurences of han- in the corpus however shows that it is actually also mutually exclusive with the plural person markers. This is clearly visible in conjuntion with the locative prefix dū-: with singular subjects, the default prefix combination in locative clauses is dū-han-; with plural subjects, han- is dropped, resulting in plain dū-dū-, dū-dɨp- and dū-ni-.
Banks' "emphatic translocative" prefix dan- can be analysed as a combination of translocative dɨ- and locative (h)an-. This is evident again from its occcurence in locative clauses, where singular -dū-d-an- contrasts with plural -dū-dɨ-ni-. Also in indicative clauses, singular -m-d-an- appears next to plural -m-dɨ-ni- (e.g. in past tense gindan- vs. gindini-).
As mentioned above, the prefix man- is distinct from the prefix ma- which is described in the next section. The element man- only appears in the prefixes daman-, guman-, cuman-, man-, duman-, naman-, gaman-, gʷaman-, which can only combine with the ablative prefix cɨ-. Most notably, these prefixes only appear with a singular subject. In corresponding clauses with a plural subject, we find the simple indicative forms -m-dū-, -m-dɨp- and -ni- instead (e.g. singular guman- against plural gindū-/gindɨp-/gini-). The most straightforward analysis is thus to decompose man- into indicative m- + locative (h)an-, with the same loss of initial h in the latter as is also observed in the portmonteau prefix dan-.
4.4. Position 2C: Cislocative ma-
The cislocative prefix ma- indicates motion towards the deictic center (comparable to German her), but can also be used in a more abstract manner indicating change of state.
The cislocative prefix cannot co-occur with the translocative prefix dɨ-, nor with the locative prefix (h)an-. Further, it does not appear in locative clauses. Otherwise, it can appear in indicative, subjunctive, temporal and focus clauses. Its position immediately following the plural person markers can be seen from several examples in the corpus, e.g.
4.5. Position 2C: Ablative cɨ-
The ablative prefix indicates movement from an anaphorical known or explicitly expressed location. It can cooccur with the locative prefix (h)an- and the plural person markers, but not with the translocative prefix dɨ- and the cislocative prefix ma-. The use of the ablative prefix in a locative clause and in conjuction with the locative prefix (h)an- is illustrated in the following examples:
5. The copula prefix ihi-/i-
In Central Kalaypuya, nouns can appear in equative clauses taking full verbal prefixation. Most equative clause in the corpus are cleft clauses, with a relative-marked verb as subject.
Nouns carrying a possessive prefix can directly take a tense-marked verb prefix when appearing in predicate position, i.e. Central Kalaypuya has a zero-copula for possessed nouns.
Non-possessed nouns have to carry the prefix an- by default. In equative clauses, the prefix an- is dropped, and the noun takes the copula prefix ihi-/i-, which is directly attached to the verb prefix.
The corresponding free nouns are an-c̓īk "yellowhammer" and a-sni "coyote".
The prefix i- has a variant ɨ- following the third person present tense prefix m-, resulting in mi-, mu- and mə-.
The exact distribution of the two suffixes ihi- and i- cannot be fully established from the corpus. In P̓īnefu, the form ī- is very common, which might represent the contracted equivalent to Santiam ihi-.
In negative clauses, the copula is -hi- or -he- in Santiam, pointing at underlying hɨ-.
[TO BE CONTINUED]
References
Banks, Jonathan. 2007. The Verbal Morphology of Santiam Kalapuya. Northwest Journal of Linguistics 1. 1-98.
Jacobs, Melville. 1945. Kalapuya Texts. Seattle: University of Washington. 394pp.
Last revised: 24 Mar 2018 (minor copy edits: 26 Dec 2020)
URL: http://ezlinguistics.blogspot.de/p/kalapuya-verb-prefixes.html
1. Introduction
The Kalapuyan languages are a small family of Native American languages that were spoken in the Willamette Valley (Oregon). Internally, the family can be divided into three branches: Northern, Central and Southern Kalapuyan. The Kalapuyan family is usually included as part of the proposed Penutian macrofamily. The closest relative of the Kalapuyan languages seems to be Takelma, based on evidence from shared lexicon and bound morphemes, although this remains controversial.
The main source for Kalapuyan languages is the large corpus of texts published by Melville Jacobs (1945). The bulk of these texts is from Central Kalapuyan: Santiam texts were collected by Jacobs himself, whereas material from the P̓īnefu (= Mary's River) and Lower McKenzie dialects were originally collected by Leo Frachtenberg in the 1910s. A first analysis of the grammar of Santiam Kalapuya was undertaken by Jonathan Banks (2007), concentrating on the verbal morphology of Santiam. This paper is based on Banks' article, and also on my own comprehensive analysis of Jacobs' text corpus. In several points, I arrive at conclusions different from Banks' results.
In Kalapuyan languages, every verb is expanded to the left by diverse inflectional prefixes, indicating tense, subordination, person/number and location/direction. Completely lacking are prefixes with "concrete" meaning (i.e. expressing body-parts, shape etc.) which are quite common in other languages of the area, including Takelma.
The concatenation of these prefixes is quite transparent and regular, with only few phonetic variations due to regular morphophonological processes. This is quite in contrast with the rather complex and "synthetic" nature of verb suffixation at the opposite side. We can therefore assume that the leftward verb affixation developed at a relatively late stage, probably through capture (via clitization) of earlier independent auxilliaries. Contact with Athabascan and Chinookan languages may have played a role in this development. Note that the presumed closest relative of Kalaypuan, i.e. Takelma, has almost no inflectional prefixes; rather, all verbal inflection for person and TAM involves highly "synthetic" suffixes.
1.1 Notes on spelling and phonology
In the text corpus collected by Jacobs, a relatively narrow transcription is employed. The most prominent example is the representation of phonemic /a/ as a or e (spelled "ɛ" by Jacobs) in the Santiam texts. I will follow Jacobs's spelling when citing examples with a or e, but in the formal description, I will always use phonemic a.
For the obstruents, I follow Jacobs' spelling of unaspirated lenes as b/d/g, but without small caps. Aspirated fortis stops are indicated by p/t/k. Affricates are respelled as c/c̓.
In many unstressed syllables, including prefixes, the contrast between short i and u appears to be non-phonemic. This can be exemplified by the past tense prefix gi-/gu-. The back variant u appears preceding a labial consonant, whereas i appears preceding alveolar consonants:
gum-, guma-
gida-, gidū-, gidī-, gini-.
gida-, gidū-, gidī-, gini-.
The prefix gum- changes to gin- when final m assimilates to an alveolar consonant in a following prefix:
gindi-, gindū-, gindan-.
The front variant i often (especially in the P̓īnefu dialect) becomes zero in open syllables, e.g.
gda-, gdī-.
The same pattern is also found with the first person plural future tense prefix dum-, the SAP present tense prefix cu-/ci-, and the third person present tense prefix um-. In the prefix cu-/ci-, the vowel is regularly dropped preceding the alveolar stop d, resulting in sda-, sdū-.
Neutralization of short high vowels is also found with the shortened variants of the first and second person plural subject prefixes dū- and dīp-: for the former, we have short du- and di-, with the latter variant dominating; dīp- is always shortened to dup-, having back u triggered by the final labial stop.
Since the variation between i and u is largely conditioned by the phonetic environment, I propose a single high central vowel phoneme here, viz. /ɨ/. The above-listed prefixes will thus be spelled:
gɨ-, cɨ-, dɨm-, ɨm-, dɨ- (shortened variant of dū-), dɨp- (shortened variant of dīp-).
No variation is found with the third person plural prefix ni-, which invariably has a front vowel.
2. Banks' analysis of Santiam verb prefixes
Banks proposes 11 prefix positions, which can be tabulated as follows:
Position 1
g-
Ø
c- ~ s-
du- ~ di-
n-
g-
d-
gi-
Ø
c- ~ s-
du- ~ di-
n-
g-
d-
gi-
Past Tense Realis
3rd Subject Non-Past Realis
SAP Subject Non-Past Realis
1st Subject Irrealis
2nd Subject Irrealis
3rd Subject Irrealis
Narrative/Habitual
Infinitive Marker
3rd Subject Non-Past Realis
SAP Subject Non-Past Realis
1st Subject Irrealis
2nd Subject Irrealis
3rd Subject Irrealis
Narrative/Habitual
Infinitive Marker
Position 2
u- ~ i-
a-
a-
Realis
Irrealis
Irrealis
Position 3
a-
Assertive
Position 4
du- ~ de-
Relative Locative
Position 5
m- ~ n-
Finite Verb Marker
Position 6
de- ~ da-
Negative
Position 7
di-
i-
i-
Subordinate Realis
Subordinate Irrealis
Subordinate Irrealis
Position 8
t- ~ d- ~ di-
de-
de-
Translocative
Emphatic Translocative
Emphatic Translocative
Position 9
du- ~ di-
dup- ~ dip-
ni-
dup- ~ dip-
ni-
1st pl Subject
2nd pl Subject
3rd pl Subject/Object
2nd pl Subject
3rd pl Subject/Object
Position 10
ma-
he- ~ ha-
he- ~ ha-
Cislocative
Proximal Deictic/Static Locative
Proximal Deictic/Static Locative
Position 11
ci- ~ ce- ~ cu-
Ablative
I mainly differ in the analysis of the prefixes appearing in positions 1 to 7. The distinction of seven positions preceding the translocative prefix d(i)‑ is over-atomistic, and some of the formative elements are in part incorrectly identified. Basically, the elements filling Banksʼ positions 1 to 3 indicate tense, partially distinguishing person, while the elements of positions 4 to 7 seven indicate syntactic function. The following section 3 will elaborate on this in detail.
The prefixes appearing in Banks' position 8 to 11 will be discussed in section 4. In Section 5, I will describe the copula prefix ihi-/i-, which primarily appears with unpossessed nouns in equative clauses. This prefix has not been described to my knowlegde so far in earlier research.
3. Macro-Position 1: tense-marking and clause relational prefixes (Bank's positions 1 to 7)
In place of Banks' positions 1 to 7, I posit one macro-position 1, which contains both mono- and poly-morphemic prefixes. Most prefixes in macro-position 1 indicate tense (out of six tenses) and clause-related function (independent clause or one out of four dependent clause types), and can be arranged into two "neat" major paradigms. Two tense-neutral prefixes appear in embedded clauses functioning as same-subject and switch-subject infinitive.
3.1. Tense-marking prefixes
Santiam Kalapuya distinguishes six tenses. Their function can be abstracted from the text corpus.
- past:
employed with remote past and irrealis events; the default tense form for narratives; also used in hypothetical clauses (in conjuction with conditional clauses headed by gī-), or in negative clauses expressing impossibility.
- habitual:
describes action performed regularly, in the past, present and future; the default tense in texts describing daily routines.
- recent past I:
recently completed actions; speaker is certain about the statement.
- recent past II:
recently completed actions; speaker is uncertain or even doubtful about the statement.
- present:
this tense covers all event which are temporally close to the present, i.e. present actions, immediate future, and very recent events which still have an effect on the present (= present perfect).
- future:
actions which will occur in the non-immediate future.
The first four tenses are expressed by person-neutral prefixes, while in present tense, third person forms contrast with common forms for first and second person (speech act participants = SAP). In future tense, there is a distinct prefix for all three persons.
The tense-relational-markers of macro-position 1 are composed of one or two prefixes and can be divided into two classes of paradigms, based on their internal structure:
Class 1: non-future tenses
Class 2: future
3.1.1. Non-future tenses
Tense-relational-markers of Class 1 can be tabulated as follows:
| independent | subjunctive | locative | temporal | relative | |
|---|---|---|---|---|---|
| habitual | dam- | dada- | dadū- | dadī- | dī- |
| recent past I | gʷam- | gʷanda- | gʷadū- | gʷadī- | gʷī- |
| recent past II | gam- | ? | gadū- | gadī- | gī- |
| past | gum- | gida- | gidū- | gidī- | gan- |
| SAP present | cum- | sda- | sdū- | sdī- | can- |
| 3 present | um- | inda- | dū- | indī- | ū- |
The non-relative forms can be analyzed in a straightforward manner:
| independent | subjunctive | locative | temporal | |
|---|---|---|---|---|
| habitual | da-m- | da-da- | da-dū- | da-dī- |
| recent past I | gʷa-m- | gʷa-nda- | gʷa-dū- | gʷa-dī- |
| recent past II | ga-m- | ? | ga-dū- | ga-dī- |
| past | gɨ-m- | gɨ-da- | gɨ-dū- | gɨ-dī- |
| SAP present | cɨ-m- | s-da- | s-dū- | s-dī- |
| 3 present | ɨ-m- | ɨ-nda- | dū- | ɨ-ndī- |
We can identify the following formative elememts:
Position A
d-
gʷ-
g-
g-
c-/s-
Ø-
gʷ-
g-
g-
c-/s-
Ø-
habitual
recent past I
recent past II
past
SAP present
3 present
recent past I
recent past II
past
SAP present
3 present
+ thematic -a-
+ thematic -a-
+ thematic -a-
+ thematic -ɨ-
+ thematic -ɨ-/-Ø-
+ thematic -ɨ-/-Ø-
+ thematic -a-
+ thematic -a-
+ thematic -ɨ-
+ thematic -ɨ-/-Ø-
+ thematic -ɨ-/-Ø-
Position B
-m-
-(n)da-
-dū-
-(n)dī-
-(n)da-
-dū-
-(n)dī-
independent
subjunctive
locative
temporal
subjunctive
locative
temporal
Prefix components in Position A indicate tense, composed of a characteristic consonant followed by a thematic vowel, while the elements in Position B indicate the function of the clause as independent or one out of three dependent clause types. As we can see, the habitual and recent past prefixes carry the same thematic vowel a, whereas the past and present prefixes agree in taking the vowel ɨ.
The prefix -da- appears as variant of locative -dū-, exclusively in the position preceding the spatial prefix han-: in the Santiam corpus, we find several instances of -de-hen- next to "regular" -dū-hen-; in the P̓īnefu texts, we regularly find contracted -dan- corresponding to Santiam -dū-hen- or -de-hen-.
Note that third person present is zero-marked with the locative prefix -dū-. The independent prefix ɨm- also has a zero-variant when it precedes a prefix element beginning with a nasal, viz. ni- "third person plural" and ma- "cislocative", e.g. ma-lami "he brings it inside", ni-dahai "they kill him".
The relative prefixes are used in relative and cleft clauses relating to subject and object (e.g. gan-hūkni "[the one] who ate it" or "[the thing] that he ate"). Like the basic tense prefixes, they are composed of the actual tense-marking consonant prefix, and a characteritic thematic element.
Relative prefixes
d-
gʷ-
g-
g-
c-
gʷ-
g-
g-
c-
habitual
recent past I
recent past II
past
SAP present
recent past I
recent past II
past
SAP present
+ thematic -ī-
+ thematic -ī-
+ thematic -ī-
+ thematic -an-
+ thematic -an-
+ thematic -ī-
+ thematic -ī-
+ thematic -an-
+ thematic -an-
We thus find the same pattern as with the basic tense prefixes: habitual and recent past agree in taking -ī- as formative element, next to past and SAP present which have -an-.
The prefixes can- and gan- are analyzed as c-a-m- and g-a-m- by Banks, contaning the "assertive" preifx a- plus the "finite" prefix m-. The final nasal is however definitely n, as can be seen from examples such as gan-hūkni. The vocalic part a- is analyzed as the same element as in the recent past prefix gʷa-. Since these two environments are actually diametrically opposite, I take -an- as an undividable relative marking component.
3.1.2. Future tense prefixes
Class 1c contains future tense markers distiguishing first, second and third person. Here, we find some signifacant differences in the paradigms of Santiam and P̓īnefu.
| Santiam | independent | subjunctive | conditional |
| 1 future | dum- | dinda- | dumi- |
| 2 future | nam- | nanda- | nami- |
| 3 future | gam- | ganda- | gami- |
| P̓īnefu | independent | subjunctive | conditional |
| 1 future | dum- | inda- | dumi- |
| 2 future | dam- | dada- | dami- |
| 3 future | gam- | gada- | gami- |
The first person prefixes largely agree, except for the loss of initial d in the subjunctive form in P̓īnefu. Loss of d is also found preceding the first person plural prefix dū-, resulting in indū-.
Second person prefixes have different characteristic consonants: Santiam has n, whereas P̓īnefu has the stop d. External evidence from Northern Kalapuyan (da-) points at *d for Proto Central Kalapuyan.
The most significant difference is found with the subjunctive prefixes for second and third person. In P̓īnefu, we find the same pattern as in the remaining tense forms, where the final nasal m found in the independent prefixes is replaced by the subjunctive marker da-.
past
2 future
3 future
independent
gɨm-
dam-
gam-
gɨm-
dam-
gam-
subjunctive
gɨda-
dada-
gada-
gɨda-
dada-
gada-
The Santiam Kalapuya forms however are characterized by having final m as an inseparable part of the future tense prefixes:
past
2 future
3 future
independent
gɨm-
dam-
gam-
gɨm-
dam-
gam-
subjunctive
gɨda-
namda-
gamda-
gɨda-
namda-
gamda-
The formative elemenents of the future prefixes can thus be tabulated as follows:
Position A
Santiam
dɨm-
nam-
gam-
dɨm-
nam-
gam-
P̓īnefu
(d)ɨm-
da-
ga-
(d)ɨm-
da-
ga-
1 future
2 future
3 future
Position B
Ø
-da-
-i-
-da-
-i-
-m-
-da-
-mi-
-da-
-mi-
independent
dependent
conditional
dependent
conditional
3.2. Tense-neutral subordinating prefixes
Next to the tense marking prefixes, Santiam has two tense-neutral subordinating prefixes. The infinitive prefix dumī‑ marks same subject complement clauses, e.g.
cum1-huli dumi1-kane
(SAP.present-want INF-cross)
"I want to cross".
(SAP.present-want INF-cross)
"I want to cross".
In Santiam, the infinitive prefix is homophonous to the first preson future conditional. In P̓īnefu, the infinitve prefix has a shortened variant dī‑, which occurs much more often than dumī‑; the first preson future conditional prefix however is unvariably dumī‑.
The subjunctive prefix gī- has two functions. First, it is used in hypothetical conditional clauses, further it appears in different-subject complement clauses:
ci gī1‑p̓īni gum1-ˀi cu dū1-he3‑lui amim ū1-gēwufit
(I SUBJ‑have past-go where LOC-LOC-many people pres.REL-gather)
"If I had it I would go where there are many people gathering."
(I SUBJ‑have past-go where LOC-LOC-many people pres.REL-gather)
"If I had it I would go where there are many people gathering."
waˀ sde1‑huli ma gī1‑kaninefaˀ
(not SAP.present‑want 2sg SUBJ‑take.across.1sg)
"I do not want you to take me across."
(not SAP.present‑want 2sg SUBJ‑take.across.1sg)
"I do not want you to take me across."
3.3. Imperative
The imperative is marked by default by the subjunctive prefix da- without any preceding tense marker.
De1-gamˀyec-aˀf!
(SUBJ-help-1sgO)
"Help me!"
(SUBJ-help-1sgO)
"Help me!"
It can cooccur with the translocative prefix di-/t- (position 2A) and the third person plural prefix ni- (position 2B).
De1-t2A-ˀi!
(SUBJ-TRLOC-go)
"Go!"
(SUBJ-TRLOC-go)
"Go!"
De1-di2A-ni2B-keni cehau!
(SUBJ-TRLOC-3pl-take.across other.side)
"Take them across!"
(SUBJ-TRLOC-3pl-take.across other.side)
"Take them across!"
With the second person plural prefix dɨp- (position 2B) and the cislocative prefix ma- (position 2C), the imperative is zero marked.
Dup2B-ˀi!
Ma2C-yī!
Du2B-ma2C-ˀi!
Ma2C-yī!
Du2B-ma2C-ˀi!
(2pl-go)
(CISLOC-return)
(2pl-CISLOC-go)
(CISLOC-return)
(2pl-CISLOC-go)
"Go (pl.)!"
"Come back!"
"Come here (pl.)!"
"Come back!"
"Come here (pl.)!"
4. Macro-position 2 (Banks' position 8 to 11)
Banks posits eight prefixes filling his positions 8 to 11: t-/d-/di-, de-/den-, du-/di-, dup-/dip-, ni-, ma-/man-, he-/hen- and ci-/ce-/cu-. I largely agree with Banks' analysis, except for three points:
- The nasal is an integral part of the prefixes hen- and den-. It assimilates to a following stop and is dropped preceding f and the sonorants m, n, l, w and y.
- The prefix den- is not a basic prefix, but a portmonteau form of di- and hen- (dɨ- and han- in my phonemic spelling).
- There are two distinct prefixes ma- and man-. The prefix ma- is invariant in all positions and also appears before stops, whereas man- has an homophonous allomorph ma- that is used in the same environment as the allomorphs he- and de-. Further, based on its distribution, man- can be analyzed as a portmonteau of the indicative prefix -m- (position 1B) and han-.
All combinations of these prefixes that are found in the text corpus can be arranged in the following grid:
Ø-
dū-
dīp-
ni-
han-
dū-
dīp-
ni-
han-
dɨ-
dɨdū-
dɨdīp-
dɨni-
dan-
dɨdū-
dɨdīp-
dɨni-
dan-
ma-
dūma-
dɨma-
nima-
---
dūma-
dɨma-
nima-
---
cɨ-
dūcɨ-
(dīpcɨ-)
nicɨ-
hancɨ-
dūcɨ-
(dīpcɨ-)
nicɨ-
hancɨ-
We can see that dɨ-, ma- and cɨ- are mutually exclusive, even though they appear in different positions. Likewise, han- and the plural person markers cannot cooccur. The combination dīp-cɨ- is unattested but this must be due to the relataively low text frequency of cɨ- in the corpus as compared to dɨ- and ma- (cf. dū-cɨ-, which only occurs once in the whole corpus). The only real gap in the grid (there is no han- + ma-) is the result of a funcional constraint against ma- appearing with any spatial-related affix, i.e. not just with han- as seen in the grid but also with the locative clause-marking prefix dū- (position 1B).
Based on this grid, I posit three sub-positions for macro-position 2: 2A and 2C for the directional prefixes dɨ-, ma- and cɨ-, and 2B for the person marking prefixes and han-.
4.1. Position 2A: The translocative prefix dɨ-
The "translocative" prefix indicates movement away from the deictic center (speaker, or the main participant in a narrative) towards a further reference point, similar to German hin.
In most cases, the prefix appears as di- in the corpus. The apocopated variant also occurs frequently, whereas the variant du- is only found preceding the second person plural marker dɨp-.
4.2. Position 2B: The plural person markers
The first and second person plural prefixes dū- and dīp- always refer to the subject of the clause, whereas the third person plural prefix nī- can refer to subject or object. Dū-, dīp- and nī- each have a shortened variant, viz. dɨ-, dɨp-, which are subject to neutralization of short high vowels in unstressed position, and ni- which invariably has the vowel i. The second person plural prefix further drops the final stop when preceding the cislocative prefix ma-, resulting in dɨ-ma-.
In present and future tense, dū- and dīp- are always bound to the corresponding tense-relational prefix:
cindu-
cindup-
dindu-
nandup-
cindup-
dindu-
nandup-
(< cɨm- + dɨ-)
(< cɨm- + dɨp-)
(< dɨm- + dɨ-)
(< nam- + dɨp-)
(< cɨm- + dɨp-)
(< dɨm- + dɨ-)
(< nam- + dɨp-)
"1pl present"
"2pl present"
"1pl future"
"2pl future"
"2pl present"
"1pl future"
"2pl future"
The prefix nī- however can coocur with first and second person tense-prefixes, in which case it expresses a third person plural object:
cini-
dini-
nani-
dini-
nani-
(< cɨm- + ni-)
(< dɨm- + ni-)
(< nam- + ni-)
(< dɨm- + ni-)
(< nam- + ni-)
"1sg/2sg subject, 3pl object present"
"1sg subject, 3pl object future"
"2sg subject, 3pl object future"
"1sg subject, 3pl object future"
"2sg subject, 3pl object future"
4.3. Position 2B: The locative marker (h)an-
The locative marker (h)an- also appears in position 2B. In Banks' analysis, the "proximal" prefix han- occupies the position immediately following the plural person markers, together with cislocative ma-. A synopsis of all occurences of han- in the corpus however shows that it is actually also mutually exclusive with the plural person markers. This is clearly visible in conjuntion with the locative prefix dū-: with singular subjects, the default prefix combination in locative clauses is dū-han-; with plural subjects, han- is dropped, resulting in plain dū-dū-, dū-dɨp- and dū-ni-.
Banks' "emphatic translocative" prefix dan- can be analysed as a combination of translocative dɨ- and locative (h)an-. This is evident again from its occcurence in locative clauses, where singular -dū-d-an- contrasts with plural -dū-dɨ-ni-. Also in indicative clauses, singular -m-d-an- appears next to plural -m-dɨ-ni- (e.g. in past tense gindan- vs. gindini-).
As mentioned above, the prefix man- is distinct from the prefix ma- which is described in the next section. The element man- only appears in the prefixes daman-, guman-, cuman-, man-, duman-, naman-, gaman-, gʷaman-, which can only combine with the ablative prefix cɨ-. Most notably, these prefixes only appear with a singular subject. In corresponding clauses with a plural subject, we find the simple indicative forms -m-dū-, -m-dɨp- and -ni- instead (e.g. singular guman- against plural gindū-/gindɨp-/gini-). The most straightforward analysis is thus to decompose man- into indicative m- + locative (h)an-, with the same loss of initial h in the latter as is also observed in the portmonteau prefix dan-.
4.4. Position 2C: Cislocative ma-
The cislocative prefix ma- indicates motion towards the deictic center (comparable to German her), but can also be used in a more abstract manner indicating change of state.
The cislocative prefix cannot co-occur with the translocative prefix dɨ-, nor with the locative prefix (h)an-. Further, it does not appear in locative clauses. Otherwise, it can appear in indicative, subjunctive, temporal and focus clauses. Its position immediately following the plural person markers can be seen from several examples in the corpus, e.g.
gidī1-ni2B-ma2C-k̓ʷa
(PAST.SUBJ-3pl-CISL-bring)
"when they brought (the woman)...".
(PAST.SUBJ-3pl-CISL-bring)
"when they brought (the woman)...".
4.5. Position 2C: Ablative cɨ-
The ablative prefix indicates movement from an anaphorical known or explicitly expressed location. It can cooccur with the locative prefix (h)an- and the plural person markers, but not with the translocative prefix dɨ- and the cislocative prefix ma-. The use of the ablative prefix in a locative clause and in conjuction with the locative prefix (h)an- is illustrated in the following examples:
gʷadū1-dū2B-ci2C-yemp
(recent.past.LOC-1pl-ABL-originate)
"at the place where we have come from".
(recent.past.LOC-1pl-ABL-originate)
"at the place where we have come from".
Cu cum1-an2B-ci2C-yemp?
(where SAP.present-CISL-ABL-originate)
"Where do you come from?"
(where SAP.present-CISL-ABL-originate)
"Where do you come from?"
5. The copula prefix ihi-/i-
In Central Kalaypuya, nouns can appear in equative clauses taking full verbal prefixation. Most equative clause in the corpus are cleft clauses, with a relative-marked verb as subject.
Nouns carrying a possessive prefix can directly take a tense-marked verb prefix when appearing in predicate position, i.e. Central Kalaypuya has a zero-copula for possessed nouns.
gʷauk gus in-din-hēduba.
(3s DIST 3.present-3s.POSS-bed)
"That is his bed."
(3s DIST 3.present-3s.POSS-bed)
"That is his bed."
gus gin-di-ˀla gan-hūkni.
(DIST past-3s.POSS-feces past.REL-eat)
"It was his [own] feces that he ate."
(DIST past-3s.POSS-feces past.REL-eat)
"It was his [own] feces that he ate."
Non-possessed nouns have to carry the prefix an- by default. In equative clauses, the prefix an- is dropped, and the noun takes the copula prefix ihi-/i-, which is directly attached to the verb prefix.
cum-ihi-c̓īk.
(SAP.present-COP-yellowhammer)
"I am Yellowhammer."
(SAP.present-COP-yellowhammer)
"I am Yellowhammer."
gʷauk gum-i-sni gus gan-t̓esdu gus dū-mā.
(3s past-COP-coyote DIST past.REL-sit DIST LOC-house)
"It was coyote himself who was there at the house."
(3s past-COP-coyote DIST past.REL-sit DIST LOC-house)
"It was coyote himself who was there at the house."
The corresponding free nouns are an-c̓īk "yellowhammer" and a-sni "coyote".
The prefix i- has a variant ɨ- following the third person present tense prefix m-, resulting in mi-, mu- and mə-.
mə-sayum u-dūleˀnafī.
(3.present.COP-grizzly DIST 3.present.REL-kill.2pl)
"It is grizzly who has been killing you."
(3.present.COP-grizzly DIST 3.present.REL-kill.2pl)
"It is grizzly who has been killing you."
The exact distribution of the two suffixes ihi- and i- cannot be fully established from the corpus. In P̓īnefu, the form ī- is very common, which might represent the contracted equivalent to Santiam ihi-.
cum-ī-ˀyēˀ. (P)
(SAP.present-COP-who)
"Who are you?"
(SAP.present-COP-who)
"Who are you?"
cum-ī-ciˀ wī. (P)
(SAP.present-COP-I)
"It is just I."
(SAP.present-COP-I)
"It is just I."
In negative clauses, the copula is -hi- or -he- in Santiam, pointing at underlying hɨ-.
wa gidē-hi-mīs.
(not past.SUBJ-COP-camas)
"It was not camas."
(not past.SUBJ-COP-camas)
"It was not camas."
wa nde-he-tmuwak.
(not 3.present.SUBJ-COP-salmon)
"It is not a salmon."
(not 3.present.SUBJ-COP-salmon)
"It is not a salmon."
[TO BE CONTINUED]
References
Banks, Jonathan. 2007. The Verbal Morphology of Santiam Kalapuya. Northwest Journal of Linguistics 1. 1-98.
Jacobs, Melville. 1945. Kalapuya Texts. Seattle: University of Washington. 394pp.
No comments:
Post a Comment