2.5. Voice.
The morphology of voice in Saanich involves a system of highly productive suffixes that function to indicate various relationships between the participants and the action or state expressed in the predicate. The transitive suffixes (§2.5.2) indicate the involvement of at least two participants: a subject and an object. See §2.4.5 for summary paradigms for subjects and objects with various transitivizers. The factive suffixes (§2.5.3) indicate the secondary involvement of a goal as well as a patient. The intransitive suffixes (§2.5.4) differ among themselves in whether the subject is agent or patient. They differ formally from the transitives in that they never occur with the object suffixes (see §2.4.3).

2.5.1. Control.
Control as a major category in the Salish languages has been discussed in a number of places (see especially Thompson, 1979b). This category cuts across the entire language with each morpheme, including each root, being either marked or unmarked with respect to control. The marked morphemes are either ‘control’ or ‘non-control’. A ‘control’ morpheme implies a participant exerting conscious control over the activity expressed in the predicate. A ‘non-control’ morpheme implies a participant expressly not exerting control. Since control involves reference to participants, it is particularly important in the voice system. In fact, this section may just as appropriately be titled "Control" and be organized in terms of ‘control’, ‘non-control’, and unmarked for control morphemes rather than in terms of transitive, factive, and intransitive.
Control, to the native speaker of English, is a semantically striking category. It seems at first that the ‘control/non-control’ distinction is one of voluntary versus involuntary action or purposive vs. accidental occurrence. However, one soon encounters sentences like 1 recorded at two different times in different contexts and having two glosses that seem to be at opposite ends of a semantic scale.
1. nəqʷnáŋət sən ‘I finally managed to fall asleep (after tossing and turning for hours trying to get to sleep)’ OR ‘I fell asleep (accidentally while driving).’

While the first gloss is clearly volitive, the second is just as clearly non-volitive. What the two glosses for example 1 have in common is that the subject fell asleep and was not able to exert control over it. Such pairs of glosses are readily and repeatedly given by informants.
Although all the suffixes of the voice system are either marked or unmarked with respect to control there are three pairs of suffixes that differ primarily in the control status they assign to the subject. The following examples briefly illustrate two of these pairs. The following sections give further exemplifications and more detailed descriptions.
2a. k̕ʷə́nət sən ‘I looked at it.’
b. k̕ʷə́nnəxʷ sən ‘I saw it.’
3a. t̕ə́m̕təl ɬtə ‘We hit each other (while throwing rocks at each other).’
b. t̕əm̕nə́kʷəl ɬtə ‘We hit each other (we ran into each other accidentally).’
4a. ʔəw̕ nəslál kʷə nəsə́q̕t ‘I intentionally tore it (a rag).’ ∥ʔəw̕ nə-s√lal kʷə nə√səq̕-ət∥ [CONTEMP 1POS-S√intend SUB 1POS√tear-CTRAN]
b. ʔəw̕ nəslál kʷə nəsə́q̕nəxʷ ‘I intentionally tore it accidentally.’ ∥ʔəw̕ nə-s√lal kʷə nə√səq̕-nəxʷ∥ [CONTEMP 1POS-S√intend SUB 1POS√tear-CTRAN]

The sentences in example 2 contain the same root and differ only in that the first has the ‘control transitive’ while the second has the ‘non-control transitive’. ‘Looking’ is a controlled activity while ‘seeing’ is not.
A gloss like that given in 4b makes one think at first that the informant is working too hard. But given the notion of control and appropriate context it makes perfect sense: ‘a nail was sticking out of the wall and I deliberately rubbed up against it and tore my shirt.’ It corresponds to the idea sometimes expressed in English as "accidentally on purpose" but without the humorous connotations.

2.5.2. Transitives.
There are three common and highly productive transitive suffixes: ∥-ət∥ ‘control transitive’, ∥-naxʷ∥ ‘non-control transitive’, and ∥-staxʷ∥ ‘causative’. Two others are quite rare: ∥-nəs∥ ‘purposive’ and ∥-əs∥ ‘effort’. The last of these may actually be an allomorph of one of the others. ∥-ət∥ ‘control transitive’. The presence of this suffix implies a patient object and an agent subject which exerts conscious control over the activity expressed in the predicate.
This suffix appears on the surface in six forms: /-ə́t/, /-ət/, /-t/, /-ə́/, /-ə/, and ø. It is stressed (examples 5-6) when following a vowelless CC root with an obstruent as the second consonant in the ‘non-actual’ (see § Otherwise the unstressed vowel is present or not (examples 7-10) depending on regular rules for /ə/ deletion. The /t/ deletes (examples 6,8,9) when followed by the /s/ of the first and second person objects (see §2.4.3).
5. sč̕ə́t sxʷ ‘You spanked him.’ ∥√sč̕-ət-ø sxʷ∥ [√spank-CTRAN-3OBJ 2SUBJ]
6. sč̕ə́s sxʷ ‘You spanked me.’ ∥sč̕-ət-s sxʷ∥ [√spank-CTRAN-1OBJ 2SUBJ]
7. k̕ʷə́nət ‘Look at it.’ ∥√k̕ʷən-ət∥
8. k̕ʷə́nəsə sən ‘I looked at you.’ ∥√k̕ʷən-ət-sə sən∥ [√see-CTRAN-2OBJ 1SUBJ]
9. léʔs sxʷ ‘You fixed me up.’ ∥√leʔ-ət-s sxʷ∥ [√repair-CTRAN-1OBJ 2SUBJ]
10. leʔtál̕xʷ sxʷ ‘You fixed us up.’ ∥√leʔ-ət-al̕xʷ sxʷ∥ [√repair-CTRAN-1PLOBJ 2SUBJ] ∥-naxʷ∥ ‘non-control transitive’. This suffix is opposed to both the ‘control transitive’ and the ‘causative’. Its presence implies a patient object and an agent subject but the subject does not exert conscious control over the activity expressed in the predicate.
This suffix appears in four forms: /-náxʷ/, /-nəxʷ/, /-ná/, and /-n/. The underlying vowel surfaces only when stressed. The suffix is stressed (examples 11-14) only when occuring with a underlyingly vowelless root (see § and no other suffixes. The form /-nəxʷ/ appears with third person object otherwise (examples 15-16). The /xʷ/ is deleted when it is followed by other suffixes (examples 13,14,17).
11. t̕əm̕náxʷ sən ‘I hit it accidentally; I finally managed to hit it.’ ∥√t̕m̕-naxʷ-ø sən∥ [√hit-NTRAN-3OBJ 1SUBJ]
12. t̕əsnáxʷ sxʷ ‘You broke it accidentally; you finally managed to break it.’ ∥√t̕s-naxʷ-ø sxʷ∥ [√break-NTRAN-3OBJ 2SUBJ]
13. t̕əm̕nás ‘He hit him (accidentally).’ ∥√t̕m̕-naxʷ-ø-əs∥ [√hit-NTRAN-3OBJ-3SUBJ]
14. t̕əm̕náŋ ‘Somebody got hit (accidentally).’ ∥√t̕m̕-naxʷ-əŋ∥ [√hit-NTRAN-PASS]
15. ʔətátnəxʷ sən ‘I finally managed to get him to sleep.’ ∥√ʔitat-naxʷ-ø sən∥ [√sleep-NTRAN-3OBJ 1SUBJ]
16. háqʷnəxʷ sən ‘I smell it.’ ∥√haqʷ-naxʷ-ø sən∥ [√smell-NTRAN-3OBJ 1SUBJ]
17. k̕ʷənnáŋəs sxʷ ‘You see me.’ ∥√k̕ʷən-naxʷ-aŋəs sxʷ∥ [√see-NTRAN-1OBJ 2SUBJ] ∥-staxʷ∥ ‘causative’. This suffix is opposed to the ‘control transitive’ and the ‘non-control transitive’. Its presence implies a patient object and an agent subject that exerts conscious control over the activity expressed in the predicate. It differs from the ‘control transitive’ in that it adds the implication that the subject is causing or obliging the activity expressed in the predicate.
The complete underlying form never appears on the surface. But the morphophonemic rules that lead to each of the five surface forms, /-táxʷ/, /-stxʷ/, /-txʷ/, /-st/, and /-t/, are straightforward. The underlying vowel appears only when the root is vowelless and the word contains no other stressable morpheme, that is, the /a/ appears only when it is the only underlying vowel in the word (examples 14-15). The underlying /s/ appears only when this suffix is preceded by a vowel (examples 16-18). The /s/ may actually represent an entirely different morpheme, but there is not enough evidence to analyze it clearly as such. As with ∥-naxʷ∥ ‘non-control transitive’, the /xʷ/ is deleted when followed by other suffixes (examples 18,19,21-23). Note that often in the ‘passive’ (§ the ‘causative’ will be phonologically identical to the ‘control transitive’. Also it is often the case that the ‘control transitive’ can be distinguished from the ‘causative’ on the surface only with reference to their different object suffix sets (see §2.4.3).
18. čəqtáxʷ sən ‘I made it big.’ ∥√čq-staxʷ-ø sən∥ [√big-CAUS-3OBJ 1SUBJ]
19. x̣əl̕táxʷ sən ‘I designed it.’ ∥√x̣l̕-staxʷ-ø sən∥ [√draw, write-CAUS-3OBJ 1SUBJ]
20. ʔət̕θíŋəstxʷ ‘Get him dressed.’ ∥√ʔit̕θ-niy-staxʷ-ø∥ [√get dressed-REL-CAUS-3OBJ]
21. štəŋístxʷ sən ‘I walked him (a dog); I ran it (a machine).’ ∥√št-ŋiy-staxʷ-ø sən∥ [√locomote-REL-CAUS-3OBJ 1SUBJ]
22. t̕ək̕ʷistáŋəs sxʷ ‘You took me home.’ ∥√t̕ak̕ʷ-ŋiy-staxʷ-aŋəs sxʷ∥ [√go home-REL-CAUS-1OBJ 2SUBJ]
23. ʔəɬənístəs ‘He fed it (the dog).’ ∥√ʔiɬən-ŋiy-staxʷ-ø-əs∥ [√eat-REL-CAUS-3OBJ-3SUBJ]
24. x̣éʔx̣əʔtxʷ sxʷ ‘You embarrassed him.’ ∥C1V́C2+√x̣eʔ-staxʷ-ø sxʷ∥ [CHAR+√embarrass-CAUS-3OBJ 2SUBJ]
25. ʔətattáŋə sən ‘I put him to sleep.’ ∥√ʔitat-staxʷ-aŋə sən∥ [√sleep-CAUS-2OBJ 1SUBJ]
26. hiθtáŋəs sxʷ ‘You kept me for a long time.’ ∥√hiθ-staxʷ-aŋəs sxʷ∥ [√long time-CAUS-1OBJ 2SUBJ]
27. híθtəŋ ɬtə ‘We were kept for a long time.’ ∥√hiθ-staxʷ-əŋ ɬtə∥ [√long time-CAUS-PASS 1PLSUBJ]

It seems that many of the stems that can occur with the ‘causative’ cannot occur with the ‘control transitive’. For all of the stems listed in this section corresponding forms with the ‘control transitive’ in place of the ‘causative’ are rejected. But see examples 53 and 55 below in § for a pair showing the difference between the two transitivizers with the same stem. ∥-nəs∥ ‘purposive’. This suffix is rare in the corpus and seems to appear on a very limited number of stems. It has been found occurring with only five different roots: √ʔax̣ʷ ‘go to’, √yeʔ ‘go, depart’, √kʷəniŋat ‘run’, √ʔənʔé ‘come’, and √št ‘walk’. Enough paradigms have been elicited, however, to show that this suffix parallels the other transitivizing suffixes. Its presence implies a patient object and an agent subject that exerts conscious control over the activity expressed in the predicate. It differs from the ‘control transitive’ in that it adds the implication that the subject has a specific purpose in its action. It often, but not always, carries an implication of malevolent intent. This suffix apparently suffers no morphophonemic variation. Since this suffix occurs no more than twenty times in the corpus the analysis here must be considered very tentative. It may turn out that this is no transitivizer at all, but another one of the factives similar to the ‘relational’ (§
28. √yéʔ-nəs sən səʔ ‘I’ll go (to do something).’ [√go-PURPOS 1SUBJ FUT]
29. √ʔáx̣ʷ-nəs sən ‘I went over there (for some specific purpose).’ [√go to-PURPOS 1SUBJ]
30. √ʔənʔé-nəs-əŋ sən ʔə tsə s√qéx̣əʔ ‘The dog came at me (probably not in a friendly way).’
31. štə́ŋnəs sən ʔal̕ ‘I’m just going to walk over there (with something in mind such as to tell somebody off).’
32. √kʷənəŋát-nəs-əs ‘He ran after it.’ [√run-PURPOS-3SUBJ]
33. kʷənəŋàtnəsáŋəs sxʷ ‘You ran after me.’ ∥-əs∥ ‘effort’? This suffix has been recorded as a clearly separable unit with only a few roots:
34. √sékʷ-əs sən ‘I put it down.’
35. √nə́w̕-əs sən ‘I put it inside.’
36. √sqé-s sən ‘I put it outside.’
37. √kʷéyl-əs sən ‘I hid it.’
38. √čákʷ-əs sən ‘I used it.’

Two of the roots have been recorded in the ‘resultive’ form with the ‘durative’ suffix and the ‘causative’ transitivizer and just about the same glosses:
39. s√náw̕-əɬ-txʷ sən ‘I got it inside.’ [STAT√inside(RES)-DUR-CAUS 1SUBJ]
40. séqəɬtxʷ sən ‘I put it outside.’

It may be that this suffix is not a transitivizer at all. There is no clear reason to separate it from the ‘purposive’ suffix, which itself has a questionable status as a transitivizer. Furthermore, there is a possibility that this suffix is to be identified with the /s/ of the ‘causative’ transitivizer analyzed in § as ∥-staxʷ∥. Examples 41-42 seem to point in this direction.
41. θɬíŋətxʷ sən ‘I stood him (a baby) up.’ (θíɬəŋ sən ‘I’m standing.’)
42. θɬíŋəstxʷ sən ‘I put it (a tent) up.’

This is, however, the only such pair recorded, and each of these was recorded only once. Example 41 is a case of the causative following a vowel without the /s/ (see § and 55 § Nevertheless, the glosses seem to differ in a way comparable to the glosses of examples 34-38.
A problem in considering this ∥-əs∥ suffix to be identical to the /s/ in the ‘causative’ is that paradigms involving ∥-əs∥ have been recorded without any trace of the expected forms of the ‘causative’. Examples 43-45 with 36 form part of a paradigm for ∥-əs∥ as a transitivizer.
43. √sqé-ŋ sən ‘He put me outside.’ [√outside-PASS 1SUBJ]
44. √sqə-ŋí-ŋə sən ‘I put you outside.’ [√outside-REL-2OBJ 1SUBJ]
45. √sqə-ŋí-ŋəs sxʷ ‘You put me outside.’ [√outside-REL-1OBJ 2SUBJ]

The /s/ of the suffix deletes when followed by another suffix as does the /xʷ/ of the ‘non-control transitive’ and the ‘causative’. See §2.4.3 on the form of the first and second person objects.
I will continue to gloss this morpheme as ‘effort’ when it is clearly not ‘causative’ or ‘purposive’. More data are needed to clarify the issue here.

2.5.3. Factives.
The factives are two suffixes, ∥-si∥ ‘indirective’ and ∥-ŋiy∥ ‘relational’, that imply a secondary involvement between the object and the subject. This involvement may be benefactive, malafactive, or neutrafactive. These suffixes occupy the same position class and appear immediately after the stem and before a transitive suffix. ∥-si∥ ‘indirective’. When this suffix is present a third participant is implied and the participant expressed in the object suffix is the goal rather than the patient of the predicate.
46. k̕ʷə́nət sən ‘I looked at it.’ ∥√k̕ʷən-ət-ø sən∥ [√see-CTRAN-3OBJ 1SUBJ]
47. k̕ʷənsít sən ‘I looked at it for him (e.g. a boat he was thinking about buying).’ ∥√k̕ʷən-si-ət-ø sən∥ [√see-INDIRECT-CTRAN 1SUBJ]
48. k̕ʷə́nəsə sən ‘I looked at you.’ ∥√k̕ʷən-ət-sə sən∥ [√see-CTRAN-2OBJ 1SUBJ]
49. k̕ʷənsísə sən ‘I looked at it for you.’ ∥√k̕ʷən-si-ət-sə sən∥ [√see-INDIRECT-CTRAN-2OBJ 1SUBJ]

In examples 46 and 48 the participant indicated in the object suffix is the patient. Examples 47 and 49 differ from these two only in the presence of the ‘indirective’ suffix. In 47 and 49 the participant indicated by the object suffix is the goal or beneficiary of the action. See § for a complete paradigm involving this root and suffix combination.
Though other glosses come to mind (such as ‘benefactive’ or ‘substitutive’), ‘indirective’ seems more appropriate. This morpheme seems to be the Saanich cognate of a suffix that has been widely noted in Interior Salishan (Carlson, 1976; Thompson and Thompson, 1976; Mattina, 1982). In Saanich, as in the Interior languages, this morpheme often seems to convey a benefactive meaning or the idea that the subject is acting as a substitute for the object as in examples 47, 49, and 50-53.
50. leʔsít sxʷ ‘You fixed it for him.’ ∥√leʔ-si-ət-ø sxʷ∥ [√repair-INDIRECT-CTRAN-3OBJ 2SUBJ]
51. leʔsís sxʷ ‘You fixed it for me.’ ∥√leʔ-si-ət-s sxʷ∥ [√repair-INDIRECT-CTRAN-1OBJ 2SUBJ]
52. x̣ətsísəs ‘He got it ready for me.’ ∥√x̣ət-si-ət-s-əs∥ [√prepare, make, do-INDIRECT-CTRAN-1OBJ-3SUBJ]
53. x̣ətsísə sən səʔ ‘I’ll make it (a paddle) for you.’ ∥√x̣ət-si-ət-sə sən səʔ∥ [√prepare, make, do-INDIRECT-CTRAN-2OBJ 1SUBJ FUT]

But the ‘indirective’ can also carry a malafactive sense as in 54.
54. k̕ʷeyəsít sən ‘I refused it to him.’ ∥√k̕ʷey-si-ət-ø sən∥ [√unable, forbid-INDIRECT-CTRAN-3OBJ 1SUBJ]

That most occurrences of the ‘indirective’ have a benefactive sense can be attributed to the fact that this morpheme is quite easily elicited with contexts involving the English "for" benefactive construction. The occurrences with malafactive readings are more difficult to elicit directly and usually appear only incidentally in running text.
Most recordings of the ‘indirective’ occur with ∥-ət∥, the ‘control transitive’ suffix. It has also been found occurring with ∥-staxʷ∥, the ‘causative’ as in example 55.
55. x̣ətsitáŋə sən səʔ ‘I’ll fix it(a paddle) for you.’ ∥√x̣ət-si-staxʷ-aŋə sən səʔ∥ [√prepare, make, do-INDIRECT-CAUS-2OBJ 1SUBJ FUT]

Although the expected /s/ of the ‘causative’ is missing in example 55, the object suffix clearly indicates ‘causative’ (see §2.4.3). Examples 53 and 55 mean very nearly the same thing. The difference lies in the fact that 55 implies that the paddle to be prepared already exists, where in 53 the paddle will be made from scratch.
The ‘indirective’ has not been recorded occurring with ∥-naxʷ∥, the ‘non-control transitive’. This is probably just an accidental gap in the corpus. ∥-ŋiy∥ ‘relational’. When this suffix is present the participant expressed in the object suffix is the goal as well as the patient of the predicate. Where ∥-si∥ ‘indirective’ with a transitivizer implies three participants (an agent subject, a goal object, and an unspecified patient), ∥-ŋiy∥ ‘relational’ with a transitivizer implies two (an agent subject and an object both goal and patient).
56. qʷéltxʷ sən ‘I talked to him.’ ∥√qʷel-staxʷ-ø sən∥ [√talk, say(RES)-CAUS-3OBJ 1SUBJ]
57. qʷélŋət sən ‘I gave him a talking to.’ ∥√qʷel-ŋiy-ət-ø sən∥ [√talk, say(RES)-REL-CTRAN-3OBJ 1SUBJ]

Example 56 with the ‘causative’ implies ‘I conversed with him’, ‘causing’ him to talk. The ‘control transitive’ in example 57 implies ‘I talked at him.’ The ‘relational’ in 57 adds the implication that the subject was scolding him, giving him a piece of his mind, or giving him advice: ‘I talked at him for him.’ As with ∥-si∥ ‘indirective’, the implication is not necessarily either benefactive or malafactive.
Besides the ‘control transitive’, the ‘relational’ has been recorded occurring with the ‘causative’ and the ‘effort’ transitivizers. See examples in § and §
The gloss for this morpheme, ‘relational’, is the gloss used by Thompson and Thompson (in press) in their discussion of the apparent Thompson Salish cognate ∥-min∥1. The term ‘relational’ can be thought of as referring to the fact that the morpheme indicates that there is a special relationship between the subject and the object other than that of agent/patient. No further significance should be given to the choice of terminology than that it is difficult to find a better label.
The /y/ of the underlying form surfaces as /y̕/ in the ‘actual’ (see §2.3.5). The suffix appears as /-ŋí/ when stressed and followed by a consonant (examples 21, 44, 45 above), /-ŋi/ when followed by a consonant and stress is in a following syllable (examples 58-59), /-ŋə/ when followed by a consonant and stress is in a preceding syllable (examples 20, 41, 42), /-ŋəč/ when unstressed and followed by a vowel (examples 60, 62), and /-ŋəy̕/ in the ‘actual’ and followed by a vowel (example 61). The /ŋ/ deletes in three environments: 1) following a sequence of two obstruents (examples 64-65), 2) following another nasal (examples 23, 63), and 3) following another velar (example 22 above). Thus other forms of ∥-ŋiy∥ include /-í/, /-i/, /-ə/, and ø. No cases of the /y/ ∼ /č/ surfacing without the /ŋ/ are evident in the corpus, but such should be expected.
58. ƛ̕čəlŋistáŋəs səʔ sxʷ ‘You’re going to sink me.’ ∥√ƛ̕č-il-ŋiy-staxʷ-aŋəs səʔ sxʷ∥ [√down-DIRECT-REL-CAUS-1OBJ FUT 2SUBJ]
59. səsiʔŋistál̕xʷ sxʷ ‘You scared us.’ ∥C1+√sey̕-ŋiy-staxʷ-al̕xʷ sxʷ∥ [RES+√scare-REL-CAUS-1PLOBJ 2SUBJ]
60. qʷə́lŋəčəɬtəŋ sən ‘Somebody scolded my kid.’ ∥√qʷəl-ŋiy=aɬ-ət-əŋ sən∥ [√talk-REL=offspring-CTRAN-PASS 1SUBJ]
61. qʷə́l̕ŋ̕əy̕ɬəŋ̕ sən ‘I’m scolding my kid.’ ∥√qʷəl-ŋiy=aɬ-əŋ sən∥ [√talk(ACT)-REL=offspring-CMDL 1SUBJ]
62. qʷə́lŋəčəɬsə sən ‘I scolded your child.’ ∥√qʷəl-ŋiy=aɬ-ət-sə sən∥ [√talk-REL=offspring-CTRAN-2OBJ 1SUBJ]
63. x̣ʷáʔaŋ̕təs ‘She was crying for him.’ ∥√x̣ʷa(ʔ)aŋ-ŋiy-staxʷ-ø-əs∥ [√cry(ACT)-REL-CTRAN-3OBJ-3SUBJ]
64. tčístəs ‘He brought it.’ ∥√tč-ŋiy-staxʷ-ø-əs∥ [√arrive-REL-CAUS-3OBJ-3SUBJ]
65. t̕kʷístxʷ ‘Get it across.’ ∥√t̕kʷ-ŋiy-staxʷ-ø∥ [√across-REL-CAUS-3OBJ]

Examples 60-62 show that the ‘relational’ suffix may precede a lexical suffix2. Example 61 shows that it need not occur with a transitivizing suffix. The translation given here for 61 is transitive though the Saanich form is not. An awkward but more literal translation would be "I’m doing some kid-scolding."

2.5.4. Intransitives.
The intransitivizing suffixes variously indicate the relationship between the subject and the predicate. They never occur with an object suffix unless they form a stem that is subsequently transitivized.

Stems are intransitive unless marked for transitivity with one of the suffixes listed in §2.5.2. Most stems are also basically ‘non-control’. Therefore, subjects of unmarked predicates are interpreted as patients as in examples 66-67.
66. √t̕əm̕=íqʷ sən ‘I got hit on the head.’ [√hit=head 1SUBJ]
67. √t̕ə́m̕ ə sxʷ ‘Did you get hit?’ [√hit QUEST 2SUBJ]

The transitive suffix provides the interpretation of the subject as a controling or non-controling agent as in 68.
68. √t̕ə́m̕-ət ə sxʷ ‘Did you hit it?’

Like the transitive suffixes, the intransitive suffixes provide a way of specifying the role of the subject as well as its control status. /-əlaʔ/ ∼ /-el̕s/ ‘structured activity’. When this suffix is present the subject is to be interpreted as agent. It seems to be similar in function and meaning to the ‘control middle’ § There are, however, several important differences. First, unlike the ‘control middle’, the ‘structured activity’ suffix does not seem to imply a necessarily controling agent (examples 69-72). Second, though this suffix is quite common, it is uncertain how productive it is. The meaning of forms with this morpheme are not entirely predictable from the meaning of the stem (examples 73, 78, 80-82). Third, this suffix contains the implication that the subject is participating in a regular, perhaps culturally defined activity, that is generally characterized in the meaning of the stem. It often conveys the notion of a habitual or characteristic activity (examples 73-82). And, fourth, the ‘structured activity’ morpheme can be followed by both lexical and transitivizing suffixes (examples 74, 87).
The two forms of the morpheme, /-əlaʔ/ and /-el̕s/ are the ‘non-actual’ and ‘actual’ (see §2.3.5), respectively. I have been able to find no morphophonemic rules to relate them.
69. pxʷəláʔ ‘The wind blows.’ (páxʷət ‘blow it (with mouth).’)
70. pxʷél̕s ‘The wind is blowing now.’
71. ƛ̕əpx̣él̕s ‘It’s scattering.’ (ƛ̕pə́x̣t ‘I scattered it.’)
72. t̕əm̕él̕s sən ʔal̕ ‘I’m just guessing.’ (t̕émət sən ‘I guessed it.’)
73. qəq̕él̕s ‘He’s a policeman.’ (qíq̕ət sən ‘I bound it up.’)
74. qəqél̕sew̕txʷ ‘Jail; police station.’
75. x̣əl̕əláʔ ‘Write.’ (√x̣əl̕ ‘draw, write’)
76. x̣əl̕él̕s ‘He’s a writer; he’s writing.’
77. šx̣əl̕él̕s ‘Pencil, pen.’
78. x̣čél̕s ‘He’s nosey.’ (x̣ə́čt sən ‘I figured it’; x̣čít sən ‘I know it.’)
79. q̕x̣él̕s sən ‘I’m clam digging.’
80. š√qək̕ʷ-él̕s ‘Pliers.’ [S,LOC√bite-STRUC]
81. x̣éʔel̕s ‘Transformer (an important Coast Salish spiritual personage).’ (√x̣eʔ ‘sacred’)
82. xʷəl̕k̕ʷél̕s ‘He’s rolling (a cigarette).’
83. wəsəláʔ tsə sqéx̣əʔ ‘The dog barks.’
84. wəsél̕s tsə sqéx̣əʔ ‘The dog is barking.’
85. ɬənq̕ʷəláʔ sən ‘I shot it (a gun, arrow, etc.).’ (ɬə́nəq̕ʷt sən ‘I shot him.’)
86. q̕ʷəŋél̕s ‘He’s borrowing (money).’
87. √q̕ʷəŋ-əláʔ-t-əŋ sən ‘He loaned me some money.’ [√borrow-STRUC-CTRAN-PASS 1SUBJ] ∥-əŋ∥ ‘control middle’. This suffix occurs on roots that are not inherently control oriented. It implies an agent subject exerting conscious control over the activity expressed in the predicate.
The term ‘middle’ for Salish cognates of this suffix is long established. It corresponds to the Indo-European notion of ‘middle voice’ insofar as it usually implies an agent acting in its own interest. But often the ‘control middle’ does not have any apparent reflexive/benefactive component to its meaning. It seems that in most cases the primary function of the ‘control middle’ is to indicate the controled agency of the subject. In general, it usually can be read as "do some X-ing" where X represents the translation of a basic ‘non-control’ root.
88. √qékʷ-əŋ sən ‘I rested; I’m resting.’ [√rest-CMDL 1SUBJ]
89. qʷə́l̕ŋ̕əy̕ɬəŋ̕ sən ‘I’m scolding my child.’ (See § example 61.)
90. √q̕ʷəčáxʷ-əŋ tsə nə√ƛ̕és ‘My stomach’s grumbling.’ [√grumble-CMDL DEM 1POS√stomach]
91. √q̕ə́p-əŋ̕ sən ‘I collected (money for a funeral).’ [√gather(ACT)-CMDL 1SUBJ] (Compare q̕pə́t sən ‘I collected it’; q̕épəɬ ɬtə ‘We got together.’)
92. √mə́t̕-əŋ sən ‘I bent.’ [√bend-CMDL 1SUBJ] (Compare mə́t̕ət sən ‘I bent it’; mə́t̕mət̕ tsə sqəlélŋəxʷ ‘The tree is bent.’)
93. √k̕ʷə́s-əŋ̕ sən ‘I’m counting.’ [√count(ACT)-CMDL 1SUBJ]
94. √qʷə́l-əŋ sən ‘I’m (doing some) cooking.’ (Compare qʷə́lət sən ‘I cooked it’; qʷə́l ‘It’s cooked.’)
95. √t̕θəsí-ŋ-əɬ sən ‘I’m pounding, hammering.’ [√pound-CMDL-DUR 1SUBJ] (Compare t̕θísət sən ‘I punched it’; xʷt̕θsás ‘get punched in the face.’)
96. √ʔəč̕=íkʷəs-əŋ sən ‘I dried (with a towel).’ [√wipe=body-CMDL 1SUBJ] (Compare ʔéč̕ət sən ‘I wiped it.’)
97. √ʔəx̣=θín-əŋ sən ‘I shaved.’ [√scrape=mouth-CMDL 1SUBJ] (Compare ʔíx̣ət sən ‘I scraped it.’)
98. √sə́w̕q-əŋ̕ ‘He’s whispering.’ [√whisper(ACT)-CMDL] (Compare skʷə́qsə sən ‘I whispered it to you.’)

There are a number of forms with /-əŋ/ for which neither control nor agency is evident:
99. √čá(ʔ)x̣ʷ-əŋ̕ ‘It’s thawing, melting.’ [√melt(RES)(ACT)-CMDL] (Compare čáʔx̣ʷət sən ‘I’m thawing it.’)
100. √ƛ̕č-íl-əŋ sxʷ ‘You sank.’ [√deep, low, down-DIRECT-CMDL 2SUBJ]
101. t̕ét̕θəŋ ‘It’s going sour.’ ∥-naŋət∥ ‘non-control middle’. This suffix follows a pattern parallel to the ‘control middle’. This suffix, however, usually appears on inherently ‘control’ stems and thus occurs less frequently. It indicates that the subject is an agent not exerting conscious control over the activity expressed in the predicate. It is most often translated "finally".
102. √ɬel-náŋət ɬtə ‘We finally got to shore.’ [√get to shore-NMDL 1PLSUBJ] (Compare ɬél ɬtə səʔ ‘We’ll get to shore.’)
103. √ʔitət-náŋət ‘He finally went to sleep.’ (√ʔitat ‘be asleep’)
104. √nəqʷ-náŋət sən ‘I finally managed to fall asleep.’ (See also example 1 above.)
105. √ʔiɬən-náŋət sən ‘I finally ate (after waiting a long time).’ (ʔíɬən sən ‘I ate.’)

This suffix has been found occurring with both the ‘control middle’ and the ‘causative’ as in examples 106 and 107, but it is rejected with ‘causative’ and other roots. More systematic elicitation is required to clear up these problems and those with the ‘control middle’.
106. √qekʷ-əŋ-náŋət sən ‘I finally managed to rest.’ [√rest-CMDL-NMDL 1SUBJ] (Compare example 88 and kʷɬ sqéqəw̕ sən ‘I’m resting now (‘stative’ and ‘actual’).’)
107. √qekʷ-əŋ-náŋət-txʷ sən ‘I finally managed to rest it (a horse).’
108. * nəqʷnáŋəttxʷ sən (with expected meaning ‘I finally managed to get him to sleep.’ Compare with example 104.) ∥-əŋ∥ ‘passive’. This suffix creates intransitive stems from transitive ones. When it occurs it always follows one of the transitive suffixes described in §2.5.2. It has the function of reversing the status of the transitive subject from agent to patient without affecting the control status. It is therefore unmarked for control.
The ‘passive’ is formally an intransitivizing suffix for two reasons. First, it is either agentless, has an indefinite agent, or the agent is marked by ∥ʔə∥, the ‘oblique’ case marker (see § as in other intransitives. When the ‘passive’ is present no object suffixes are. This is one of the defining characteristics of the intransitives. And, second, in most transitives the third person subject is marked ∥-əs∥ (see§2.4.4). In intransitives third person subject is ø. The ‘passive’ patterns with the intransitives and takes a ø third person subject.

Although the ‘passive’ and the ‘control middle’ are both intransitive suffixes and are phonetically identical, they must be recognized as two separate morphemes. Their functions are quite different: ‘control middle’ implies a controlled agent subject, where ‘passive’ is unmarked for control and implies a patient subject.
The following examples illustrate the ‘passive’ occurring with various subjects and each of the transitivizing suffixes.
109. qʷə́lŋəčəɬtəŋ sən ‘Somebody scolded my kid (I was child-scolded).’ ∥√qʷəl-ŋiy=aɬ-ət-əŋ sən∥ [√talk-REL=offspring-CTRAN-PASS 1SUBJ]
110. ʔətáttəŋ sən ‘It put me to sleep; I was put to sleep.’ ∥√ʔitat-staxʷ-əŋ sən∥ [√sleep-CAUS-PASS 1SUBJ]
111. híθtəŋ ɬtə ‘We were kept for a long time.’ ∥√hiθ-staxʷ-əŋ ɬtə∥ [√long time-CAUS-PASS 1PLSUBJ]
112. kʷənátnəsəŋ ‘Somebody’s running after him; he’s being run after.’ ∥√kʷəniŋat-nəs-əŋ∥ [√run-PURPOS-PASS]
113. sqéŋ sən ‘Someone put/took me outside; I was put outside.’ (Compare with examples 36, 43-45 above)
114. čákʷəŋ sən ‘They used me; I was used (employed).’ ∥√čakʷ-əs-əŋ sən∥ [√use-EFFORT-PASS 1SUBJ] (See § on the deletion of the /s/.)
115. k̕ʷənstəŋ sxʷ ‘He showed it to you (you were caused to see it).’ ∥√k̕ʷən-staxʷ-əŋ sxʷ∥ [√see-CAUS-PASS 2SUBJ]
116. k̕ʷənnəŋ sxʷ ‘Someone saw you; you were seen.’ ∥√k̕ʷən-naxʷ-əŋ sxʷ∥ [√see-NTRAN-PASS]
117. wələwástəŋ tsə swə́y̕qəʔ ʔə tsə sqéx̣əʔ ‘The man got barked at by the dog.’ ∥C1(l)+√was-ət-əŋ tsə s√wəy̕qəʔ ʔə tsə s√qex̣əʔ∥ [RES(PL)+√bark-CTRAN-PASS DEM S√man OBL DEM S√dog]
118. ɬə́tx̣təŋ̕ sən ‘I’m shivering.’ ∥√ɬətx̣-ət-əŋ sən∥ [√shake(ACT)-CTRAN-PASS 1SUBJ]
119. ʔəx̣θínətəŋ sən ‘I got shaved.’ (Compare with example 97 above.)
120. x̣əl̕ástəŋ sən kʷəʔ ‘He took my picture; I got x-rayed.’ ∥x̣əl̕=as-ət-əŋ sən kʷəʔ∥ [√write, draw=face-CTRAN-PASS 1SUBJ INFORM]
121. x̣əl̕ásnəŋ sən kʷəʔ ‘I got my picture taken (I was in the picture by chance).’ ∥√x̣əl̕=as-naxʷ-əŋ sən kʷəʔ∥ [√write, draw=face-NTRAN-PASS 1SUBJ INFORM]
122. x̣ɬə́təŋ sən ‘Somebody hurt me.’ ∥√x̣ɬ-ət-əŋ sən∥ [√feel bad-CTRAN-PASS 1SUBJ]
123. x̣əɬnáŋ sən ‘I got (my feelings hurt).’ ∥√x̣ɬ-naxʷ-əŋ sən∥ [√feel bad-NTRAN-PASS 1SUBJ] ∥-tal∥ ‘control reciprocal’. This suffix creates a formally intransitive stem with two implied participants. Each participant is both agent and patient. It is most often translated into English as "each other". This suffix additionally implies that both participants exert conscious control.
124. sə́y̕p̕təl̕ ‘They’re tickling each other.’ (səč̕ə́p̕s sxʷ ‘You tickled me.’)
125. sə́y̕p̕təl̕ ɬtə ‘We’re tickling each other.’
126. sə́y̕p̕təl̕ sxʷ ‘You’re tickling each other.’

The subject need not be plural, and forms with no overtly marked subject are ambiguous between implied first and third person subjects.
127. ʔəwáʔtəl̕ sən ‘I’m racing.’ ∥√ʔəwá(ʔ)-tal sən∥ [√contest(ACT)-CRECIP 1SUBJ]
128. t̕ə́m̕təl ‘We hit each other; they hit each other.’ (Compare with example 136 below.)
129. ʔəx̣θínətəl ‘We shaved each other; they shaved each other.’ (Compare with examples 97 and 119 above.)

Specific participants can be indicated in two ways: 1) with the addition of transitive and object suffixes (example 131), and 2) with the specification of an oblique object (examples 132-133).

130. kʷíw̕ən̕təl̕ ‘They’re fighting.’ ∥C1V́+√win-tal∥ [ACT+√fight-CRECIP]
131. kʷìw̕ən̕təl̕táŋ̕əs ‘We’re fighting; he’s fighting with me.’ ∥C1V́+√win-tal-staxʷ-aŋəs∥ [ACT+√fight-CRECIP-CAUS-1OBJ]
132. kʷíw̕ən̕təl̕ sən ʔə tsəw̕níɬ ‘I’m fighting with him.’
133. kʷíw̕ən̕təl̕ sən ʔə tɬ nə́kʷə ‘I’m fighting with you.’

In examples 131-133 both the subject and object are equally agent and patient. The subject, however, is marked ‘control’ by the ‘control reciprocal’ suffix so that, in this context, the implication is that the subject is the instigator of the fight. See also examples 142-145 below for object and subject markings on the ‘non-control reciprocal’.
The underlying /a/ of the ‘control reciprocal’ surfaces, like the underlying vowels of the ‘control transitive’, ‘non-control transitive’, and the ‘causative’, only when the root is vowelless and no other stressable vowels are present.
134. x̣əčtál ɬtə ‘We know each other.’ ∥√x̣č-tal∥ [√figure, know-CRECIP 1PLSUBJ]

It may be that this suffix is further analyzable as involving the /t/ of the ‘control transitive’. The fact that the difference between a ‘control’ and ‘non-control’ transitive is often carried by the difference between /t/ and /n/ suggests the separation of the /t/ and /n/ from the ‘control’ and ‘non-control’ reciprocals. Since there is no supporting evidence for such an analysis and nothing lost or gained either way, this description will consider the /t/ and /n/ parts of the morphemes.3 ∥-nəwel∥ ‘non-control reciprocal’. This suffix functions just like the ‘control reciprocal’ except that it implies that neither participant is exerting conscious control. As with the ‘control reciprocal’ it is often translated with "each other," and the subject need not be plural. When no overt subject is mentioned the form is ambiguous between first and second persons.
135. pítnəkʷəl ‘We recognized each other; they recognized each other.’ (pítnəxʷ sən ‘I recognized him.’)
136. t̕əm̕nə́kʷəl ‘We hit (ran into) each other; they hit each other.’ (Compare with example 128 above.)
137. k̕ʷənnə́kʷəl ɬtə ‘We saw each other.’
138. xʷčəm̕snə́kʷəl sən ‘I met him (on the road, not for the first time).’ ∥xʷ√čəm̕=as-nəwel sən∥ [LOC√meet=face-NRECIP 1SUBJ]
139. nəɬnə́kʷəl ‘They look like (are) each other.’ ∥√niɬ-nəwel∥ [√it is-NRECIP]

As with the ‘control reciprocal’ specific participants can be mentioned. But the ‘non-control reciprocal’ differs in that one of the two mechanisms available to the ‘control reciprocal’ for this purpose is not used with the ‘non-control reciprocal’. Specific participants are indicated only by the addition of transitive and object suffixes. It is in this form that the underlying /e/ surfaces. It takes stress when foll owed by a single syllable (see §1.4 on stress placement).
140. ʔəθnəkʷélət sən ‘I spliced it (rope).’ ∥√ʔaθ-nəwel-ət-ø sən∥ [√front-NRECIP-CTRAN-3OBJ 1SUBJ]
141. ƛ̕əq̕nəkʷélət sxʷ ‘You pressed it together.’ ∥√ƛ̕iq̕-nəwel-ət-ø sxʷ∥ [√press-NRECIP-CTRAN-3OBJ 2SUBJ]

It should be noted that in examples 140-141 the subject and object are not the primary participants in the reciprocal activity. In 140, for example, it is the two facing ends of the rope.
The /w/ appears only when it is glottalized by the ‘actual’ resonant glottalization (see §2.3.5). In the following examples the forms are in the ‘actual’ and contain subjects and objects that are the primary participants in the reciprocal activity.
142. qʷə̀l̕qʷəl̕n̕əw̕əl̕táŋ̕ə sən [CHAR+√talk(ACT)-NRECIP-CAUS-2OBJ 1SUBJ]
143. qʷə̀l̕qʷəl̕n̕əw̕əl̕táŋ̕əs sxʷ [CHAR+√talk(ACT)-NRECIP-CAUS-1OBJ 2SUBJ]
144. qʷə̀l̕qʷəl̕n̕əw̕əl̕tál̕xʷəs [CHAR+√talk(ACT)-NRECIP-CAUS-1PLOBJ-3SUBJ]
145. qʷə̀l̕qʷəl̕n̕əw̕əl̕tál̕xʷ sxʷ [CHAR+√talk(ACT)-NRECIP-CAUS-1PLOBJ 2SUBJ]

Examples 142-145 all have the same translation: ‘We were talking, chatting together.’ Example 142 implies ‘I’m telling you something;’ 143 implies ‘you’re telling me something.’ Examples 144-145 imply that more than two people are chatting. The former excludes, while the latter includes, the addressee. ∥-sat∥ ‘reflexive’. This suffix creates formally intransitive stems where a single participant, the subject, is implied and is both agent and patient. It is often translated with English "self".
This morpheme is unmarked for control. A control/non-control contrast for reflexives can be produced with the inclusion of ∥-ət∥ ‘control transitive’. Since most roots are ‘non-control’, the addition of ‘reflexive’ to a root produces a ‘non-control reflexive’. The addition of the ‘control transitive’ to the root before the ‘reflexive’ has the effect of producing a ‘control reflexive’.
Since the /t/ of ∥-ət∥ ‘control transitive’ always deletes before the /s/ of the ‘reflexive’ as it does before the object suffixes (see §, the presence of the transitivizer can often be seen only in the meaning of the form. There are, however, forms where ∥-ət∥ shows itself before the ‘reflexive’. In § it was shown that some roots are best considered to be underlyingly vowelless. When these roots are in the ‘non-actual’ form stress falls on the vowel of the suffix. The underlying vowels of several suffixes (‘control transitive’, ‘non-control transitive’, ‘causative’, and ‘control reciprocal’) surface only when following these vowelless roots. This is also true of the ‘reflexive’. Examples 146 to 148 show the /a/ of the reflexive following vowelless roots.
146. √ʔiʔ-sát yəq sxʷ ‘I hope you’ll get better.’ [√good-REFL OPT 2SUBJ]
147. √čəq-sát ‘It’s getting bigger.’ [√big-REFL] (Compare /čəqnáxʷ sən/ ‘I made it (a sweater) big (accidentally);’ /čəqtáxʷ sən/ ‘I made it big (meaning to).’)
148. kʷɬ iʔ √pəq̕-sa(ʔə)t ‘She’s getting gray (hair).’ [REAL ACCOM √white-REFL(ACT)]

Notice that each of these involves a ‘non-control’ predicate.
There are some cases, though, where the ‘reflexive’ follows a vowelless root and the underlying /a/ does not surface. Examples 149 and 150 illustrate this.
149. tčə́sət sən ‘I poked myself.’
150. θk̕ʷə́sət ‘He stretched himself.’ (Compare /θk̕ʷnáxʷ sən/ ‘I managed to straighten it;’ /θk̕ʷə́t sən/ ‘I straightened it.’)

Notice that these involve ‘control’ predicates.
Since roots like ∥√θk̕ʷ∥ ‘poke’ and ∥√čq∥ ‘big’ pattern together with respect to other suffixes, we would expect them to pattern together here. That is, the expected surface form of ∥√θk̕ʷ∥ with ∥-sat∥ would be */θək̕ʷsát/. Since a semantic difference between forms with /-sát/ and forms with /-ə́sət/ is one of ‘non-control’ versus ‘control’ we might assume two separate morphemes: a ‘non-control reflexive’ and a ‘control reflexive’. But another difference between forms exemplified in 146-148 and those in 149-150 is that the latter are logically, though not formally, transitive; an action mediates an agent and a patient. This all suggests that forms with /-ə́sət/, involving both semantic categories of control and transitivity as well as reflexivity, should be analyzed as involving two suffixes: ∥-ət∥ ‘control transitive’ and ∥-sat∥ ‘reflexive’. The /a/ of the ‘reflexive’ does not surface because the stress is taken by the preceding suffix. The analysis for examples 149 and 150 will then be as follows: ∥√tč-ət-sat∥ and ∥√θk̕ʷ-ət-sat∥. The reduction of /a/ to /ə/, the deletion of the ‘control transitive’ /t/ before /s/, and the placement of stress on the /ə/ of the ‘control transitive’ are well justified elsewhere (see §1.4, §, §2.4.3, and §
Following are a few more examples of the ‘reflexive’.
151. √šém-əsət sxʷ ‘You dried yourself (in the sun).’
152. mə́l̕č̕sət sən ‘I’m rolling over.’
153. ɬə́məx̣ʷsət ‘He rubbed himself.’
154. məxʷísət ‘He rocked (himself in a chair or boat).’
155. ƛ̕pə́x̣sət ɬtə ‘We scattered.’
156. həlísət ‘He saved himself (his soul).’ (√həlí ‘alive’)
157. čə́qʷsət sən ‘I have a fever.’ (√čəqʷ ‘fire’. Not ‘I burned myself’ or ‘I set myself on fire.’4)
158. čéw̕sət ‘He’s showing off.’ (čéw̕t sən ‘I’m admiring him.’)

159. ƛ̕áməsət ‘It fits (a coat).’ (√ƛ̕am ‘correct, well’)
160. x̣éƛ̕sət ‘It’s a storm.’ (√x̣eƛ̕ ‘windy’)

Since there are at least three other ways in Saanich to express reflexive-like ideas (‘control middle’, ‘non-control middle’, and the bare ‘non-control’ root), the number of occurrences of this suffix in the corpus is quite limited.
Notes to §2.5.

1. Shifts of Proto-Salish *n to i have been noted in Interior Salish. See especially Kinkade (1982). This suffix in Saanich illustrates a case to the expected intermediate stage of *n > y.

2. These examples are complex forms exhibiting a secondary layer of derivation. While not unusual in Saanich and Salishan in general, secondary derivation is not common. This important topic in Saanich morphology is not covered in this work, but it deserves further investigation.