Chapter 17: Voice: Passive, Antipassive, and Focus

Starting point#

A Tz'ikin is visiting his parents' house in Tecpán and admires the tortillas and a new blouse.

Look at the verbs: xb'anon "made" (who made?), xb'an "was made", nponon "pats (tortillas)", xkemon "wove", xkem "was woven", xk'ayïx "was sold", xloq' "was bought", xtoj "was paid". Not one of them has the familiar transitive shape x-u-b'än, x-u-kem. Each has shed one of its two participants. This chapter explains how and why.

Here is the story of that blouse as a short text. Read it for the gist now and again after "How it works".

Re po't re' xkem pa Iximche' roma ri te Ixk'at. Rija' jantape' nkemon. Xkem pa jun ik'. Jeb'ël rub'onil: käq, q'än chuqa' räx. Xk'ayïx pa k'ayb'äl pa Chiq'a'l. ¿Achike xloq'on? Jun ixöq aj Tz'olöj Ya' xloq'on. Xtoj k'iy pwäq roma ri po't. Wakami ri ixöq nukusaj ri po't pa nimaq'ij. Ri te Ixk'at nub'än jun chik wakami; ri xta Ixchel nrajo' ri k'ak'a' po't. "Ja rin xiloq'on ri b'ätz'," nub'ij ri xta Ixchel, "po ja ri nute' xkemon."

This blouse was woven in Tecpán by te Ixk'at. She is always weaving. It was woven in a month. Its colors are beautiful: red, yellow and green. It was sold at the market in Comalapa. Who bought it? A woman from Sololá bought it. A lot of money was paid for the blouse. Now the woman wears the blouse at the fiesta. Te Ixk'at is making another one now; xta Ixchel wants the new blouse. "It was I who bought the thread," says xta Ixchel, "but it was my mother who wove it."

How it works#

Why voice exists#

A transitive verb (Chapter 11: Transitive Verbs in the Incompletive) carries two participants on the verb itself: the doer, or agent, marked by Set A (ergative), and the thing done to, or patient, marked by Set B (absolutive):

x-Ø-u-kem  ri po't  ri te Ixk'at
COMPL-3sB-3sA-weave  the blouse  the te Ixk'at
"te Ixk'at wove the blouse"

Voice is the grammatical term for a change in the verb that rearranges these participants. Kaqchikel has two such changes, and both turn the transitive verb into an intransitive one with a single participant marked by Set B:

• Passive: the patient becomes the subject; the agent is dropped or pushed out into a side phrase ("the blouse was woven [by te Ixk'at]").
• Antipassive: the agent stays as subject; the patient is dropped, incorporated, or pushed into a side phrase ("te Ixk'at weaves [blouses]").

Why would a language want this? Three reasons, and they apply to both voices:

1. You do not know, or do not care to name, one participant. "My money was stolen" (who by? no idea); "she weaves" (what? whatever comes).
2. You want to keep talking about the same participant. A story about a blouse keeps the blouse as subject: it was woven, it was sold, it was bought. A story about a weaver keeps the weaver as subject: she weaves, she sells, she earns.
3. The grammar forces it. In Kaqchikel, when you ask "who did it?" or say "the man who bought the cow", the verb must be antipassive. This is not a stylistic choice; a transitive verb is simply ungrammatical there.

The reason behind reason 3 goes back to the ergative pattern of Chapters 10 and 11. Intransitive subjects and transitive objects share Set B; transitive agents stand apart with Set A. Several Mayan languages, Kaqchikel among them, do not allow the Set A agent to be pulled out of its clause for questioning, focusing or relativizing. To get at the agent, the verb is first made intransitive so that the agent becomes a Set B subject; then it can be questioned or fronted freely. That is the antipassive's most important job.

The passive of root transitives: tense the vowel#

Root transitives are one-syllable stems of the shape consonant-vowel-consonant with a lax vowel (-b'än, -löq', -tz'ët, -tük, -k'ïs). Their passive is made by a single change: the lax vowel becomes tense. Nothing is added.

Why does tensing the vowel work as a signal? Recall from Chapter 1: Sounds and Spelling that lax vowels live only in a word's final syllable. The lax vowel of a transitive root is therefore the mark of a bare, unsuffixed root; historically the passive had a suffix that pushed the root vowel out of final position and tensed it, and after the suffix wore away, the tense vowel alone was left carrying the meaning. A few roots, such as -tik "plant" and -tij "eat", are listed with a tense vowel already, so for them, as for the glottal-final roots below, the pronouns alone show the voice.

Now compare full forms. The passive is intransitive, so there is only a Set B pronoun, and it refers to the patient. Here is one pair taken apart:

x-Ø-u-b'än
COMPL-3sB-3sA-make
"she made it"

x-Ø-b'an
COMPL-3sB-be.made
"it was made"

Set B stays in the passive and Set A disappears. Here is the full set of contrasts:

Some examples in sentences:

Xb'an ri samaj.
The work was done.

Xtik ri awän pa juyu'.
The corn was planted on the hillside.

Xch'aj ri tzyäq iwir.
The clothes were washed yesterday.

Xmes ri jay. The house was swept.

Nk'is ri wa'in.
The meal is being finished. (incompletive passive)

Xixtz'et pa k'ayb'äl.
You all were seen in the market.

Glottal-final roots. Roots that end in a glottal stop (-to', -tzu', -sa', -su', -ya', -ti') already have tense vowels, so vowel change cannot mark the passive. Only the pronouns show the difference:

Listen for the Set A pronoun: if it is there (-u-, -in-, -qa-), the verb is active; if the aspect prefix and Set B sit directly on the root, it is passive.

The passive of derived transitives: -Vj becomes -Vx#

Derived transitives end in a vowel plus j (-tz'ib'aj, -k'utuj, -q'etej, -kanoj, -samajij). Their passive replaces -Vj with -Vx, and the vowel of the suffix goes lax (it is now the final syllable, which is where lax vowels belong):

More examples:

Xeleq'äx ri nupwäq.
My money was stolen.

Xtz'apïx ri ruchi' jay.
The door was closed.

Xsipäx jun wuj chwe.
A book was given to me.

Xkamisäx ri kumätz.
The snake was killed.

Yek'utüx ri tijoxela' jantape'.
The students are always being asked (questions).

A third set of derived transitives comes from positional roots with the suffix -Vb'a' (-tz'uyub'a' "seat someone"); their passive is -Vb'äx (xtz'uyub'äx "he was seated"). You will meet these in Chapter 18: Positionals.

The spreading -Vx passive on roots#

Because the -Vx passive is so regular, many speakers extend it to root transitives too, adding a lax copy of the root vowel plus x: xchapäx "it was grabbed" beside xchap, xtz'etëx "it was seen" beside xtz'et, xmuqüx "it was buried" beside xmuq, xloq'öx beside xloq'. Speakers from areas with fewer vowel distinctions favor the suffixed form, since the tense/lax contrast is weak or absent for them; everywhere the most common verbs (xb'an, xloq', xtz'et) tend to keep the old vowel-change passive. Learn the vowel-change forms as primary and recognize the -Vx forms when you hear them.

Naming the agent: roma#

If you want to mention the doer in a passive sentence, add the relational noun -oma "because of, by" with the right Set A prefix (Chapter 21: Relational Nouns in Depth): woma "by me", awoma "by you", roma "by him/her", qoma, iwoma, koma. Many speakers say ruma for roma.

Xkem ri po't roma ri te Ixk'at.
The blouse was woven by te Ixk'at.

Xato' woma.
You were helped by me.

Xloq' ri wakx koma ri achi'a' aj Tz'olöj Ya'.
The cow was bought by the men from Sololá.

Xtz'ib'äx ri wuj roma ri a Tz'ikin.
The letter was written by Tz'ikin.

Compare the three ways of reporting the same event, all correct:

When to use the passive#

• Unknown or unimportant agent: Xeleq'äx ri nuch'ich'. "My car was stolen."
• Tact, avoiding blame: Xk'is ri kape. "The coffee got finished." (rather than naming who drank the last of it)
• Keeping the topic: Re po't re' xkem pa Iximche'. Xk'ayïx pa Chiq'a'l. Xloq' roma jun ixöq aj Tz'olöj Ya'. "This blouse was woven in Tecpán. It was sold in Comalapa. It was bought by a woman from Sololá." The blouse stays the subject throughout.
• Formal and narrative registers: announcements, stories and ritual speech use passives freely.

The antipassive: forms#

The antipassive keeps the agent and removes the patient from the verb. It too is intransitive, with a Set B subject, but now the Set B subject is the doer.

Root transitives add -on. Because a suffix follows, the root's lax vowel tenses (it is no longer in the final syllable):

Derived transitives replace the final j of -Vj with n:

Inflection is plain intransitive. One form taken apart:

y-i-tz'ib'an
INCOMPL-1sB-write.ANTIP
"I write"

The set: yitz'ib'an, yatz'ib'an, ntz'ib'an, yojtz'ib'an, yixtz'ib'an, yetz'ib'an; and for a root, yikemon, yakemon, nkemon, yojkemon, yixkemon, yekemon.

Many familiar nouns are built on these stems with the agent suffix -el (Chapter 7: Nouns: Plurals, People Words, and Word-Building): tijonel "teacher" (one who teaches), k'ayinel "seller", tikonel "planter", b'ixanel "singer". That is a good way to remember which stems exist.

Reintroducing the patient: richin#

A demoted patient can be mentioned in a side phrase with the relational noun -ichin "of, with respect to": wichin, awichin, richin, qichin, iwichin, kichin.

Nkemon richin ri po't ri te Ixk'at.
Te Ixk'at is weaving (at) the blouse.

Xojto'on kichin ri ak'wala'.
We helped (with respect to) the children.

Yatz'eton wichin.
You are looking at me.

This side-phrase patient is rather bookish, and it belongs first of all to the activity uses of the antipassive (Uses 1 and 2 below). In everyday speech, if you want to name a specific patient you usually just use the transitive verb. In the contexts where the grammar forces the antipassive (Uses 3 to 5), the patient usually stays in the sentence as an ordinary noun phrase, though careful speakers can use a richin phrase there too, as you will see.

Use 1: activity without an object#

The plainest use of the antipassive says what someone does, as an activity, habit or profession, without naming any particular object:

Nkemon ri te Ixk'at. Te Ixk'at weaves.

Yitz'ib'an ronojel aq'a'.
I write all night.

Nk'ayin ri xta Ixchel pa k'ayb'äl.
Ixchel sells at the market.

Yetikon ri achi'a' pa juyu'.
The men plant on the hillside.

Yojch'ajon chwa'q.
We will do washing tomorrow.

Use 2: object incorporation#

When the patient is a general kind of thing that defines the activity, it is placed right after the antipassive verb as a bare noun, with no article and no agreement on the verb. English does something similar when it compounds: to babysit, to fruit-pick.

Nkemon po't ri te Ixk'at.
Te Ixk'at weaves blouses. (She is a blouse-weaver.)

Nb'anon xajab' ri achin.
The man makes sandals.

Yojtz'ib'an wuj.
We write letters.

Ntikon kinäq' ri ma Kawoq.
Ma Kawoq plants beans.

Contrast: Nukem ri po't "she is weaving the blouse" (a specific blouse, transitive) versus nkemon po't "she weaves blouses" (what she does).

One verb form, two different jobs#

Before the last three uses, pause on a distinction that saves you from an apparent contradiction. In Uses 1 and 2 the antipassive genuinely demotes the patient: it is either gone altogether, folded in as a bare noun (never with ri), or pushed out into a richin phrase. Grammars call this the absolutive or incorporating antipassive, and for it the no-article rule is strict.

In Uses 3 to 5 the same verb form does a different job, which grammars call agent focus. Here the verb goes intransitive not to demote the patient but because the agent has been pulled out for questioning, relativizing or focusing, and Kaqchikel will not pull a Set A agent out of a transitive clause. The patient is not demoted at all: it stays in the sentence as an ordinary noun phrase, article, demonstratives and all, as in ¿Achike xkemon re po't re'? "Who wove this blouse?" The verb agrees with the questioned or focused agent, and the patient in this construction is normally a third person. You will also meet the patient of a focus clause wrapped in a richin phrase, especially in writing; the authors' sequel volume clefts that way: Ja ri Ixnal ri xkemon richin ri jeb'ël pa's "It's Ixnal who wove the beautiful sash" (Kik'aslemal ri Kaqchikela', draft, p. 225). Either shape is good Kaqchikel; this book practices the plain noun phrase.

So when you meet an antipassive verb followed by ri plus a noun, do not conclude that the incorporation rule has been broken. Ask what the antipassive is doing there: if it names an activity, the object must be bare or absent; if it lets a questioned, relativized or focused agent escape its clause, the object keeps its article.

Use 3: asking who did it#

To question the agent of a transitive verb you must use the antipassive. Achike "who, what" stands in front, the verb is intransitive and agrees with the questioned agent:

¿Achike xb'anon?
Who did it?

¿Achike xkemon re po't re'?
Who wove this blouse?

¿Achike xloq'on ri wakx?
Who bought the cow?

¿Achike xk'ayin ri ichaj?
Who sold the vegetables?

¿Achike xeleq'an ri pwäq?
Who stole the money?

If instead you question the patient, the verb stays transitive, because the agent is still there in Set A:

¿Achike xub'än?
What did she do?

¿Achike xukem ri te Ixk'at?
What did te Ixk'at weave?

¿Achike xulöq' ri achin?
What did the man buy?

The pair is worth memorizing as a unit: ¿Achike xb'anon? "Who did it?" versus ¿Achike xub'än? "What did he do?" The presence or absence of -u- is the whole difference.

Use 4: "the person who ..."#

A relative clause (a clause that describes a noun, as in "the man who bought the cow") works the same way. When the described noun is the agent inside the clause, the clause verb goes antipassive; when it is the patient, the verb stays transitive. Compare the pairs:

Notice ri ... ri: the first ri is the article of the noun, the second introduces the clause.

Use 5: focus with ja and fronted agents#

Kaqchikel normally puts the verb first and the agent last (verb, object, subject). Moving a noun in front of the verb gives it emphasis, and the strongest emphasis uses the focus word ja (Chapter 6): ja ri te Ixk'at "it is te Ixk'at". When the focused noun is the agent of a transitive, the verb goes antipassive, for the same reason as in questions and relatives: the agent has left its clause.

Ja ri te Ixk'at xkemon ri po't. It was te Ixk'at who wove the blouse.

Ja ri ma Kawoq xtikon ri awän. It was ma Kawoq who planted the corn.

Ja rin xib'anon.
It was I who did it.

Ja rije' xeloq'on ri jay.
It was they who bought the house.

Patients can be focused without any change to the verb: Ja ri po't xukem ri te Ixk'at "It was the blouse that te Ixk'at wove."

A little history: the focus antipassive and modern word order#

In older Kaqchikel, any agent placed before the verb, focused or not, triggered the antipassive. A sentence like Ri te Ixk'at xkemon ri po't "Te Ixk'at wove the blouse" was the normal way to start with the agent. Under pressure from Spanish, subject-verb-object order has become common, and the preverbal agent is no longer felt as emphatic. So today you will constantly hear fully transitive verbs with the agent in front:

Ri te Ixk'at xukem ri po't.
Te Ixk'at wove the blouse.

Ri ma Kawoq xutik ri awän.
Ma Kawoq planted the corn.

Both patterns are good Kaqchikel. The cost of the modern pattern is ambiguity when either noun could be the doer: Ri ala' xuto' ri achin can mean "the boy helped the man" or "the man helped the boy", and only context decides. The old focus antipassive had no such ambiguity, and it survives, solid as ever, in the three places where it is required: questions about the agent, relative clauses on the agent, and ja focus on the agent.

Patterns at a glance#

Stem derivation

One event three ways

When the antipassive is required

In the first three rows (the agent-focus uses) a patient noun phrase keeps its article: ¿Achike xkemon re po't re'? Only the incorporated object of the last row must be bare.

Vocabulary#

Usage and dialect notes#

• Two passives, one meaning. Xtz'et and xtz'etëx both mean "it was seen". Use the vowel-change form for the common roots, and do not be surprised by -Vx on rarer ones or from speakers of western towns. Neither is wrong.
• A third way to lose the agent: -täj. Beside the passive, Kaqchikel has a middle voice in -täj (after a consonant -itäj) meaning "get X-ed", used when no agent is assumed at all, or to sidestep blame. The authors' sequel volume contrasts all three: Ri jöb' xuyöj ri nukematz'ib' "the rain ruined my computer" (transitive), Xyoj "it was ruined" (passive), Xyojtäj "it got ruined" (middle), and asks the everyday question with it: ¿Achike xb'anatäj? "What happened?", literally "what got made?" (Kik'aslemal ri Kaqchikela', draft, pp. 228-29). This book does not drill the form; simply recognize -täj as "got X-ed, ended up X-ed".
• roma and ruma. The relational noun of agent appears as roma in Tecpán and Comalapa and as ruma in other towns; the first-person form is woma or wuma accordingly. The same word means "because of" (roma ri jöb' "because of the rain") and "because" as a clause linker (Chapter 22).
• Do not translate English passives one for one. English uses the passive to be vague about people ("mistakes were made"); Kaqchikel does too, but it also has the antipassive and the third person plural for that: xkeleq'aj ri pwäq "they stole the money" is a common way to report a theft by unknown persons.
• The antipassive is not a "general present". English speakers sometimes assume yitz'ib'an is the habitual and nintz'ib'aj the progressive. Not so: both can be habitual or ongoing. The difference is whether a specific patient is tracked on the verb.
• Pronunciation of -on. Because the root vowel tenses, -löq' sounds noticeably different in xloq'on (two clear full vowels) from the lax ö of xulöq'. Listen for it; the contrast is real and helps hearers.
• Pitfalls for English speakers:
• Asking "who did it?" with a transitive verb. ¿Achike xub'än? will be understood as "what did he do?". Who-questions about the agent need the antipassive: ¿Achike xb'anon?
• Keeping Set A in a passive. Xub'an ri samaj is not a passive; the passive has no agent pronoun at all: xb'an ri samaj.
• Forgetting that the Set B pronoun now points at the patient: xixtz'et means "you all were seen", not "you all saw".
• Leaving the root vowel lax in the antipassive: xloq'on, never xlöq'on; a suffix follows, so the vowel tenses.
• Using an article with an incorporated object: nkemon po't "she weaves blouses", not nkemon ri po't. In a plain activity sentence, "the blouse" needs the transitive: nukem ri po't. The rule stops at the agent-focus uses, though: in ¿Achike xkemon re po't re'? the article is exactly where it belongs.
• Thinking the preverbal agent always needs the antipassive. Today Ri te Ixk'at xukem ri po't is normal; the antipassive is obligatory only for questioned, relativized and ja-focused agents.

Looking back and ahead#

The passive and antipassive are the payoff of the ergative pattern of Chapters 10 and 11, and they rest on the root/derived distinction and the tense/lax vowel rule from Chapter 11: Transitive Verbs in the Incompletive and Chapter 1: Sounds and Spelling. The agent question ¿achike x-...-on? returns in Chapter 20: Asking Questions, and the relational nouns -oma and -ichin get their full paradigms in Chapter 21: Relational Nouns in Depth. Chapter 18: Positionals continues the theme of derivation: a class of roots that build statives, intransitives, transitives (and their -Vb'äx passives) and adjectives by regular suffixes.