Marianne Mithun_Contact and Time_2007

8/12/2019 Marianne Mithun_Contact and Time_2007

1/24

Journal of Language Contact THEMA 1 (2007)www. jlc-journal.org

GRAMMAR, CONTACT AND TIME

Marianne Mithun

University of California, Santa Barbara

1. The Area: California

North America is home to considerable linguistic diversity, with 55-60 distinct language families

and isolates in the traditional sense, that is, the largest genetic groups that can be established by the

comparative method. Of these, 22 are represented in California. Some California languages are

members of large families that extend over wide areas of the continent, such as Uto-Aztecan, Algic

(Athabaskan-Wiyot-Yurok), and Athabaskan-Eyak-Tlingit. Others belong to medium or small

families spoken mainly or wholly within California, such as Utian (12 languages), Yuman (11),

Pomoan (7), Chumashan (6), Shastan (4), Maidun (4), Wintun (4), Yokutsan (3), Palaihnihan (2),

and Yuki-Wappo (2). Some are isolates: Karuk, Chimariko, Yana, Washo, Esselen, and Salinan.Most of the currently accepted genetic classification was established over a century ago (Powell

1891). Locations of the languages can be seen in Figure 1.

In 1903, Dixon and Kroeber noted structural resemblances among some of these families and

isolates and ascribed them simply to common typology. Ten years later, they hypothesized that the

similarities might be due to distant genetic relationship and proposed, primarily on the basis of

grammatical structure, two possible stocks or groups of families, which they named Hokan and

Penutian.

The originalHokan hypothesislinked Karuk, Chimariko, Yana, Shastan, Palaihnihan, Pomoan,

Esselen, and Yuman. Over the next half-century, various scholars added the Washo, Salinan, and

Chumashan families from California; the Tonkawa, Karankawa, and Coahuilteco languages from

Texas; Seri, Tequistlatecan, and Tlappanec from Mexico; Subtiaba from Nicaragua; and Jicaquefrom Honduras. Some of the proposals remain promising, and others have since been abandoned.

Evaluation of their validity remains difficult, however. The proposed relationships would be

remote, comparable to those of Indo-European or deeper, and documentation of the languages, most

of which are no longer spoken, is in many cases limited. There has, furthermore, been a long history

of contact.

The original Penutian hypothesis grouped together the California families Maidun, Wintun,

Utian (Miwok-Costanoan), and Yokuts. Subsequent proposals added Takelma, Coos, Siuslaw,

Alsea, Klamath-Modoc, Cayuse-Molala, Sahaptian, Takelma, Kalapuya, and Chinookan of Oregon;

Tsimshianic of British Columbia; and Mixe-Zoquean and Huave of Mexico. Ongoing work

continues to uncover promising lexical resemblances among some of the families, but in general the


2/24


3/24

146 Marianne Mithun


(1) Central Pomo verbs

a. y:q know

da-y:q recognize by feeling around (with hands)

-y:q recognize by touching (with fingers)p

h-y:q recognize visually (by sight)

qa-y:q recognize (food) by tasting

s-y:q recognize (a drink) by tasting

ba-y:q follow an order

b. : stick together, be alongside of each other

da-: push on something that sticks in your hand

-: stick on with fingers, as chewing gum under a table

ma-: step on a nail or something that sticks in your foot

a-: sit on a thorn, put a patch on pants

h-: stick up a pole, pitchfork, shovel, etc. in the ground

m-: catch fire

ph

-: hammer a nail into the wall, nail something onp

ha-: something floating downriver gets stuck on the bank

s-: while one is drinking, something gets into the mouth that doesnt belong: bug,

dirt

a-: stick a support, as a box, next to something long, like fenceposts stored upright for

use

c. h-kn close a door

-kn close a gate

ph-kn nail something shut, like a box, blow something shut like a window

a-kn block, as a doorway

-kn close something up with the fingers

h-kn be blocked by vegetation, like a tunnel overgrown with brush

m-kn be constipated

d. da-kl brush away dirt

ma-kl wipe feet

h-kl wipe off, as a babys dirty face

s-kl wash off

-kl clean, as pebbles out of beans before cooking them

-kl pull up weeds out of lawn

ph-kl hoe dirt to make it look nice

e. -ql finish a task

p-ql finish sewing something

a-ql finish peeling, as apples or potatoes

qa-ql finish eating something

s-ql finish drinking something

ba-ql finish a speech

ph-ql finish reading a book

h-ql win a game

da-ql steal (Coast and Yorkville dialects)

A substantial proportion of the verbs in the language contain prefixes of this type.

The Central Pomo prefixes are clearly derivational. Over the course of development of the

language, speakers have used them to create new lexical items as needed for nameworthy concepts.

Speakers know which combinations exist in the language, which could exist but do not, and which

would not be well formed. The prefix h-, for example, seen above in h-knclose a door, occurs

in large numbers of Central Pomo verbs, contributing such meanings as moving a long object


4/24

Grammar, contact and time 147


lengthwise, poking, thrusting, jabbing, and throwing. The root qlfinish, complete forms

the basis for a number of verbs, as seen above: finish a task, finish sewing something, finish

peeling something, finish eating something, finish drinking something, finish a speech, finish

reading something, and more. There is, however, no verb h-ql complete poking. Apparentlyspeakers have not felt the need for such a term. The prefix-root combinations are not constructed

anew each time they are used: they are formed once, then learned, stored, and used as single lexical

items. In fact the root ql-no longer occurs on its own in the language: it now appears only in the

verbs with prefixes. The lexicalization of the prefix-root combinations is also apparent in the

translations provided by speakers. The verb -kl is literally clean by pulling, but the speaker

rendered it as weeding a lawn. The verb-klis literally clean with fine finger action, but it was

translated as picking pebbles out of beans before cooking. Not surprisingly, many derived stems

have somewhat idiomatic meanings, either because they were first coined for a specific purpose, or

because their original meanings have shifted over time. The verb ba-y:q in (1)a is literally orally

recognize, but it is usually used to mean follow an order. The verb da-ql in (1)e is literally

finish off by pulling, but in the Point-Arena/Manchester and Yorkville dialects, it is used to mean

swipe or steal.

Such morphemes have sometimes been referred to as instrumental affixes, because they often

evoke an instrument: with the hands, with the feet, with a knife, by fire, etc. They have also

been termed lexical affixes because their meanings seem quite concrete, more like those of roots

or whole words. Their meanings are actually more abstract and diffuse than those typical of nouns

denoting instruments like hand or knife. Even from the few examples above, we can see that the

prefix da-could be translated by hand, by touching, with the palm, or by pushing: da-y:q

recognize by feeling around (with hands), da-: push on something that sticks in your hand,

da-kl brush away dirt. The prefix s-, seen above in s-y:qrecognize (a drink) by tasting and s-

: while one is drinking, something gets into the mouth that doesnt belong, such as a bug or

dirt, appears in verbs of drinking, eating mushy food, sipping, sucking, smoking, swallowing,

whistling, whispering, rain, and more. The prefix a-, in a-knblock, as a doorway, a-:sit on a thorn, put a patch on pants and a-ql finish peeling, as apples or potatoes, is used in

verbs involving a massive object, especially the buttocks, but also in verbs involving cutting. The

prefix h-, seen above in

h-kl wipe off, as a babys dirty face,

h-ql win a game, and

h-kn

be blocked by vegetation, like a tunnel overgrown with brush, has an especially wide range of

meanings involving round masses like rocks and the head, shooting, gambling, flowing, and

vegetative growth. The ranges of meaning are due to several kinds of factors, only some of which

can be recovered.

One is the status of the morphemes as derivational prefixes and the lexicalization of the resulting

verb stems. Speakers have created new stems by analogy to existing ones, extending the meanings

of the prefixes in various directions. The prefix qa-, which appears in qa-y:qrecognize (food) by

tasting and qa-ql finish eating something, can indicate biting and chewing as in these two

verbs, but it also appears in verbs meaning catch, trap, prune, use pliers, lock, and others.

The semantic thread is easy to see in this case.

Others are the result of sound change. The Central Pomo prefix m-appears on the one hand in

such verbs as m- kick, m-liw knock over with the elbow or foot, m-s sniff at

something, m-l miss when trying to kick, or miss in marbles, and on the other in m-hw

unravel, m-n be clogged, as a hole clogged with dirt, m-kn be constipated, m-liw

begin to ache, m-wchap, m-yblister, m--wfeel hot, m-:catch fire,m-liw

melt, m-w be cooked, m-yl cook various things together in the same pot, and m-b:

burst open, as from swelling from heat or frost. Cognates of the prefixes occur in all languages of

the Pomoan family, as laid out by Oswalt (1976: 16). In Central Pomo, all vowels in initial syllables

except for a have disappeared. Thus Proto-Pomo*ba-, *a-, *da-, *ma-,*pha-,*qa- and *a-


5/24

148 Marianne Mithun


correspond to Central Pomo prefixes of the same shape, but both Proto-Pomo *mi- (involving

protuberance near end of long object) and *mu-(internal force, fire, heat, cold, light, emotions,

mind) correspond to modern Central Pomo m-. The phonological merger has resulted in some

doublets: the verb m-d:n means either dancing along or burning along. Proto-Pomo *i-(pulling, breathing) and *u-(hooking, dangling) both appear in Central Pomo as *-. The first

(pulling) was seen above in -klpull up weeds out of lawn. The second (hooking) was seen in

-kn close a gate. The Central Pomo vowel loss led to further phonological changes. When the

initial consonants of prefixes came into direct contact with the initial consonants of roots, they

sometimes underwent further reduction. Proto-Pomo *di-(gravity, falling; multiple long objects)

and *du-(work, finger action) first collapsed to *d-, then simply to pre-glottalization in Central

Pomo (represented here by an apostrophe ). A descendant of *di-appears in -lyakill off all of

something, such as weeds. A descendant of *du- appears in -y:q recognize by touching,

-: stick on with fingers, -kl clean, as pebbles out of beans before cooking, and -ql

finish a task. Here, too, the merger of forms has resulted in some polysemous stems. The verb

-n:wan is used both to describe someone staggering around, like a man so inebriated that he isincapable of walking and keeps falling down (falling), and for throwing a ball around, from

person to person (fine finger action).

The prefixes also differ from nouns syntactically: they do not function as syntactic arguments

and are not referential. Verbs containing them are akin to English verbs like slap, kick, poke,

bash, or burn. They may imply the involvement of a kind of entity: the verb slap usually

implies a flat surface, particularly the palm of the hand; kick implies action with the foot; poke

implies use of a long, pointed instrument; bash conjures up a heavy, blunt instrument; burn can

imply fire. But none of them introduces a referent into the discourse. I would not say He bashed me

on the head and it broke, to mean that the rock crumbled (unless the rock was already under

discussion). Such verbs can, of course, co-occur with independent referential nouns. I can say He

bashed me on the head with the rock and it broke . In a similar way, Central Pomo prefixed verbs

can co-occur with independent nouns, a further indication that the prefixes and nouns are servingdifferent functions. Furthermore, different nouns can appear with the same prefix, and vice-versa, as

in (2).

(2) Central Pomo prefixed verbs with nouns

a. Qhabwi

hb:.

qhab=wi h-b:-

rock=with with.massive.object-split-SEMELFACTIVE.PERFECTIVEHe cracked it open with a rock.

b. Qhabwi h-b:.

qhab=wi h-b:-

rock=with thrusting-split-SEMELFACTIVE.PERFECTIVE

He threw a rock (and hit someone in the head).

Although the Central Pomo prefixes most often evoke a kind of instrument (means) or manner of

action, there is no explicit specification of the role of that entity in the situation beyond general

involvement. There are relics of a few other prefixes that serve to characterize the central

participant (theme) in the situation, though these have not been productive for some time. A prefix

ba-, for example, appears in a few common verbs involving multiple participants: m-w(one) is

lying, ba-w (more than one) is lying; hmw (one) is sitting, ba-mw (more than one) is

sitting.

The Central Pomo prefixes thus typically contribute more concrete meanings than we normally

expect of grammatical affixes, but they are not equivalent semantically or syntactically to nouns or

verbs.


6/24



3. Means and manner beyond Pomoan

Cognates of the Central Pomo means/manner prefixes occur in all of the Pomoan languages. Their

ancestors are easily reconstructed as prefixes for their common parent, Proto-Pomoan. Affixes of

this type are relatively rare among languages of the world. Yet they are surprisingly common

among languages of California and adjacent areas.

3.1 The Hokan hypothesis

Prefixes with similar meanings appear in a number of other families and isolates in the

hypothesized Hokan stock: languages of the Yuman-Cochiti and Palaihnihan families and in the

isolates Karuk, Yana, and Washo.

Yumanlanguages are spoken in Southern California, western Arizona, and northern Sonora and

Baja California. Means/manner prefixes are somewhat less transparent semantically in these

languages than in Pomoan, but a number have been identified. Describing Mesa Grande Diegueo,

spoken near modern San Diego, Langdon reports:

The bulk of Diegueo verb stems are formed by the addition to a root of one of a variety of

prefixes. While the meanings of these prefixes cannot be covered by a single descriptive term,

they fall mostly in the category of instrumentals or causatives. Similarly, the meaning of a

particular prefix cannot always be clearly established, especially if it occurs with roots not

attested in other stems. (1970: 80).

Miller (1990: 32-49) provides a detailed description of their counterparts in the closely-related

Jamul Diegueo, citing cognates in Mesa Grande and Yuma, a member of a different branch of the

Yuman family. She mentions aa- motion along a path, long motion, motion with long object,

motion with instrument; ch-with the mouth or by talking, or involving multiple small objects or

repetition; k-speaking, action on foot, force; m-states; n-willful or self-assertive behavior;p-

involving pressure, s- contact with a surface, sh-with the hand or by means of a hand-held

instrument; andx-involving nature. Each of these prefixes appears in just as many verbs without

consistent or discernible meaning as with them, however. Other prefixes occur in the same position

without obvious means/manner meanings, in fact often with no detectable meaning at all. Miller

provides several sets of verbs sharing the same stems. Based on the root kwin, for example, she lists

aakwin wrap, amkwin twist, apkwin cross legs, shemkwin pinch, skwin put on or wear

earring, and terkwin wind a watch. Based on the root tu hit, strike are kuuttu kick, kuutu

pound acorns,puuturun over, stupick up several, shttushove, andxtuwind blows.

TheKaruklanguage is an isolate spoken in Northern California along the Klamath River. Bright

(1957: 86-87) describes what he characterizes as

[] a phenomenon which is marginal to normal derivation, namely the occurrence of certain

sub-morphemic elements at the beginning of verbs. These phonemic sequences, analogous to

English sn-in sniff, snort, sneeze,etc., are so limited in distribution as to make it inadvisable to

place them on the same level as the usual type of morpheme. However, a common meaningmay be traced throughout the various occurrences of each element. (1957: 86).

The sequences he lists are im- involving fire or heat, pa- with the mouth, ak- with thehand, ima- with a striking implement, ta-with an implement, vu-with a cutting implement,a- with the teeth, and im- by rubbing. Verbs containing the element pa- with the mouthinclude p-upto kiss, pa-nutto suck on, p-pu to chew, pa-snikto blow a whistle, andpa-xutto hold in ones mouth. Verbs containing ta-with an implement include t-sirto brush,t-tuyto sweep, ta-xvavto stir soup, and ta-xvuk to hook.Haas (1980) isolated several more:at- popping, if- sticking together, ik- by hitting, with blows, ik- making a sound, pa-

completely, iv-whooshing, wafting, and vim-with vigor. Examples of ik-by hitting are ik-as to mash, ik-mar to meet, hit with the fist, ik-napto nail, ik-nat- in rock to crack acorns


7/24

150 Marianne Mithun


with, ik-nik- in maul for driving wedges, ik-niv to knock, especially on oak trees, ik-pak to

chop, ik-patto break (by hitting with something), and ik-rakto split (wood or the like).

Not far from Karuk to the east are the Palaihnihanlanguages Achomawi and Atsugewi. Talmy

(1972: 407-427) describes the semantics of the Atsugewi means/manner prefixes in detail. The set

is rich and generally corresponds to those in the Pomoan languages in meaning but not usually

form. They include tu- hands working toward each other; tasting, ci- hands working

manipulatively, ma- stepping, kicking, involving foot or leg, ti- involving buttocks, pu-

involving mouth or mouth-shaped object, as a flower, wi- with teeth, pri- sucking, tasting,

smelling, phu- spitting, blown breath, hi- involving whole body, ce- involving eye or eye-

shaped object, as a button or hailstone uh- swinging, batting, throwing, pounding, chopping,

hailing,, cu- pushing with stick, pool-cueing, prodding, poking, piercing, skewering, supporting

with a cane, holding pinned against a wall, flowing liquid, downward pressure as snow on limb,

falling, car in collision, sunshade on cradleboard, tightly packed material as caulking, ra-thrusting

up, digging, awling, sewing, propping, leaning, poling, raking, sweeping, scraping, whittling,

plowing, smoothing over, hugging, being pinned down by a log, slicing, sawing, carting, driving,getting run over, pressure as from gas in stomach, ice under soil, leaning cradleboard, a shingle, ih-

/uh- planar object flush against a surface, as a spread out blanket, board nailed to a wall, ta-

paddling (hot rock around in soup), stirring, ka- boring, raining, ru- dragging with a cord,

flexing forearm, suspending with cord, binding, girding, steady pull as stream on an anchored cloth,

attached object as sinew, a belt, an icicle, mi- cutting with blade of knife, ca-wind blowing,

miw-/mu:-heat or fire, wu:-light shining, sa-/su-/si-/siw-visually, ka-/ku-/ki-/kiw-auditorily,

tu-by touch, cri-linear objects in parallel, as hairs in a plait, stalks in a sheaf, sticks in a bundle.

An example is woswalctashe threw the clothes into the laundry tub, consisting of w-uh-swal-

ictwaFACTUAL-throwing-limp.material.move-into.liquid-FACTUAL.

Immediately to the southwest of Atsugewi is Yana, extensively documented by Sapir. The

language contains a large set of elements expressing instrumentality. The Yana dictionary (Sapir

& Swadesh 1960) lists such markers as baa- striking, punching, bu- kicking, stepping, bui-involving feet, di- shearing, peeling, ga- speaking, uttering, ha- involving long object, hi-

with the hands of generally instrumentality, pulling, bending, tearing, breaking, washing, eating

with fingers, nis-by or at the foot,pu-blowing, sucking, with the mouth, wa-involving a long

object, yul- mashing, a- eating, biting, and u- shooting, digging, wind blowing. Their

functions can be seen by comparing sets of verb stems containing -balla-push, knock: baa-balla-

to push, knock down with fist, bu-ballaa to kick, bui-balla- to kick, da-balla- to nudge,

haa-balla- to strike, and uu-balla- to push. A sample of verb stems all containing the prefix ga-

speaking includes ga-cgai- to talk loudly, ga-laa- to cry, weep, ga-sapai to answer, ga-wi- to shout to, gaa-can to make a speech, ga-baari-to stop one from uttering, ga-buisdi-tospeak for ones happiness, ga-damci-to meet in council, ga-dunindi-to leave word behind, ga-

duukab- to stop talking, ga-keerailaugu- to speak in a slow drawl, ga-kuuwi- to shout to

ones dream spirit, ga-mici-to prevaricate, ga-milkuidibil-to talk unintelligibly, wrongly, ga-

rii-to speak Northern Yana dialect, ga-taa-to speak a non-Northern Yana dialect, and ga-wau-

to talk to (1960: 79).

Washois an isolate spoken to the south, around Lake Tahoe in California and Nevada. Jacobsen

describes a set of over 125 prefixes of concrete meaning (1960: 85). Some indicate a type of

means or instrument, with as de- with hand, fist, instrument held in hand, ug-with long object

swung sideways, and se- by fire or heat. Some indicate manner, such as w-by treading on, eb-

by scratching, pinching, squeezing with fingers, le- biting, holding in teeth, d- shooting,throwing rock, h

u- by pushing, stirring with side of long object, liw- by pressing, mashing,

stomping on, and bv- cutting with sawing motion. Of course means and manner are often

indistinguishable. Use of the prefixes can be seen in d-emwith.instrument.held.in.hand-find =


8/24



to find by digging, cuw-yu hanging-oscillate = to swing, and ug-:kmwith.long.object.swung.sideways-sever = to cut with axe. Some of the prefixes characterize the

entity involved, as in g-wkaw surface-be.hard = to be hard, as wood or bread, wege-pilpil

(liquid) to be blue, and l-lsil(cloth) to be thin, fine. More will be said about the Washo systemlater on.

The Chumashanlanguages, spoken on the Central Coast of California, were once hypothesized

to be possible members of the Hokan group, but the proposal was based on minimal lexical

evidence which was subsequently rejected. Nevertheless, these languages show verbal prefixes with

functions similar to those seen so far, indicating means, manner, medium, and characteristics of

entities involved in the situation. These prefixes, which immediately precede the root, vary greatly

in productivity and can have meanings that are difficult to isolate. Among the Ineseo Chumash

prefixes listed by Applegate (1966: 346-369) are api- of, with fire, aputi-of liquid in motion,

aq-with the mouth, aqul-of, with a long thin object, ani-with the feet, maq-of, with a line

or rope, naq-with the body, palof, with a pliable object, tak-with the hand, taya-of rain,

ux-of, with fire, u-with the hand, palm, wala-of, with the body or a massive or bulky object,wa-with the hand, wati-of a flow, liquid in motion,yul-of, with heat, andyuq-of the legs.

Uses of aq-/ax-with the mouth can be seen in ax-tapwith.mouth-enter = to put into the mouth,

ax-klawith.mouth-crack = to shell acorns with the teeth, ox-loqwith.mouth-be perforated = tobit a hole in, and oq-lokinwith.moith-cut = to bite in two, bite through.Others indicate manner,

such as apti-crushing, grinding, aqni-mentally, aqi-calling, aqu-eating, aqutii verbal or

mental activity, axuti-sewing, ci-following, iku-pressure, restraint, itzx-hearing, kal-by

cutting, kina- dying, kupal- drinking lu- growing, max- draggling, sliding, qal- tying,

binding, tal- grasping, holding, ti- speaking, intention, uqti- throwing, uxmal- washing,

cleaning, and wi- by blows, by hitting. Uses of this last prefix can be seen in wi-eqhitting-

split = to split by blows, wi-sqhitting-do.firmly = to pound tight, and wi-su-kitwonhitting-CAUSATIVE-exit = to knock out, dislodge with a blow.

Although sets of means/manner prefixes appear in numerous languages grouped as Hokan, notall members of the group show them. They are not mentioned in descriptions of Shasta, Esselen, or

Salinan. Such prefixes may of course have once existed and been lost, or they may simply have

escaped documentation, which in some cases is scant. The pattern is not confined to the Hokan

group, however.

3.2 The Penutian hypothesis

Sizeable sets of prefixes with the same kinds of meanings, though not the same forms, also occur in

languages of California and Oregon hypothesized to belong to the other major stock, Penutian.

TheMaidunlanguages Maidu (Northeastern), Konkow (Northwestern), and Nisenan (Southern)

were spoken at contact immediately to the south of Yana and Atsugewi and to the north of Washo

in northeastern California. All contain sets of means/manner prefixes, which tend to show vowel

harmony with the following stem. Describing Maidu proper, Dixon (1911: 693-699) lists the

following prefixes: ha(n)-with the shoulder or back, hi:-with nose or snout, i:n-by sitting on,

is-/ic-with the foot, ka-with the flat hand or flat side of something, ki-with the fingers, o:-

with the head, so:-with the arms, te-with or upon the foot, ba-/be-/bo-/k-,/bu-involving a

rounded or massive thing, ho-/hu-with the edge of a long thing, generally by a steady, continuous

motion, si-/s- with the end of a long thing, wa-/we-with the edge or side of a thing, often by a

sudden motion or blow, wo-/w-/wu-with the end of a long thing, generally by a blow, ya-/ye-

/yo-/y-/yu- with the end of a long thing, endways, or in a direction parallel to the length of the

thing, he-by accident or spontaneously, and wi-by force, especially by pulling. Some examples

of of the use of o:-with the head are o-bttodombreaking a stick with the head, o-po-pokdom

shaking water out of the hair, :-puldonto root up (as a hog), to dig up with horns, :-moto to


9/24

152 Marianne Mithun


put heads together, and :-miton to look into a house, i.e., stick head down in through smoke

hole. Some examples of wa-/wa-with the edge or side of a thing, sudden motion or blow, are wa-

pdauto scrape off with edge of a knife, wa-btonto break a stick by blow with the butt of an

axe or with a club, w-daston to split with an axe, w-hapkinto insert a stick into a bunch ofsticks, w-katsitoto bat across, knock across, with side of pole, w-kuttodombiting in two with

teeth, w-pitinto pinch with thumb and fingers, w-tsapdaudomtearing off with teeth. Similar

forms can be seen in Konkow (Ultan (1967) and Nisenan (Uldall & Shipley 1966).

Just north of the Palaihnihan languages, in Oregon, is the Klamath language, described by

Barker (1964: 114-119) and discussed in detail by DeLancey (1988, 1991, 1996). Klamath contains

72 morphemes that characterize a type of instrument, a manner of action, or a type of entity acted

upon. Among them are n- acting with a round instrument, l- acting with the fingers,

fingernails, v-acting with a long object, cin- acting with the back, go- acting with the headfirst, qb

v- acting with the mouth, sucking, spitting, kt-hitting, kicking, spi-dragging, and noy-

burning. Examples of their use are qb-oyamnawith.mouth-carrying.or.holding = has something

in the mouth, swi-tlaby.tying-squash = cinches a horse, and no-wangaburning-fall = treeburns down. Many of the morphemes in this group characterize the entity centrally involved (the

theme), such as i-acting on plural objects, q- acting on a flat object, klv- acting upon fire, andsl-to act upon a clothlike object, as in sl-wlgathrows down a cloth.

Prefixes indicating types of instruments and involved entities also occur in the Takelma

language, spoken immediately to the west of Klamath and to the north of Shastan in Oregon. Sapir

(1922a: 64-86) lists sets of prefixes, many evoking body parts and other instruments, such as i:-

hand, with the hand, waya-knife, da-mouth, burning, glowing, dakh- head, with the head, on

top of,da:- ear, with ear, hearing, s:in- nose, in nose, with nose, gel- breast, mentally, sal-

foot, with foot, al- face, with eye, seeing, looking, xa:- cutting, splitting, breaking, di:-

crushing, mashing, squeezing, di:-involving long object, la-bursting, ripping open. Examples

include han-waya-swilswlhi he tore him open with a knife, dakh-da-ha:lnda I (with.mouth)

answered him, xa:-be-nkhwan I warm my back with.the.sun. The same prefixes cancharacterize either an instrument or entity acted upon: xa:-bem-kwo:tkwidin I broke it with a

stick, be:m-wai:toxxin I gather sticks. When both are mentioned in a verb, the one

characterizing the object precedes that characterizing the instrument.

Still further north in northwest Oregon, Washington, and Idaho, is the Sahaptian family,

consisting of just two languages, Nez Perce and Sahaptin. Both contain prefixes indicating means,

manner, or the type of entity involved. In his grammar of Nez Perce, Aoki (1970: 84-86) lists 167

prefixes he terms adverbial. Among them are c:-with pointed object, c:ye-by shaking, him-

with mouth, ke- with teeth, l:w pertaining to fish, l:w in speech nikth- dragging, nim-

with eyes, n:xcwith nose,ptipounding, punching, qi-with sticky matter, qism-in anger,

sep: wind or air, sep:- blowing, su:ye- pushing, t: by speech, t:ke- with cane-like

object, tul:-with foot, we-with chopping instrument, w:-shooting, lein talking, ands-

with knife (plural objects). Examples are t:ke-keykse I am limping with a cane and ew-yetwikceI shoot as I chase it, I chase it shooting. Jacobs (1931: 157-181) provides a substantial

list for Sahaptin.

Means/manner prefixes occur in many of the languages grouped into the Penutian stock, as we

have seen, but not in all of them. There is no mention of them in descriptions of the Wintun, Utian

(Miwok-Costanoan), or Yokuts languages, nor in descriptions of the Oregon Coast languages Coos,

Siuslaw, and Alsea, nor further afield.


10/24



3.3 Wappo-Yuki

Two remotely related languages comprising a third small family were spoken in areas directly

adjacent to the Pomo. Wappo territory is immediately to the southeast of the Pomo, and Yuki

territory immediately to the north. Both Wappo and Yuki show small sets of prefixes which

characterize means, manner, and sometimes the entity involved in ways strikingly similar to those

in other languages in the area. The forms in the two languages appear to be cognate.

Among the Wappoprefixes, Radin (1929: 29) lists m-with the hand, na- with the mouth,p-

with the foot, wi- by pulling, by force, general instrumentality, and hu-head, all of which

appear in the same position in the verb. Describing m- with the hand, he notes, This is

apparently an old instrumental prefix and in only a few cases is the stem employed as an object.

(1929: 33). Among his examples are m-tspki he grabbed (with hand), m-k: to feel withhand, mai-m-wlhe scours, m-wtito catch with hand, me-ktche tickles (with fingers)me-mmahe picks up long object with hand, me-mnumahe picks up flat object with hand, me-k:luma he picks up concave object with hand, andm-ptcskihe folded his arms. In his Wappo

dictionary, Sawyer (1965) contrasts meh-wyi:si rub with the hands, rub by hand and pih-wyi:sirub with the foot. Some additional verbs with the prefixphe/pih-involving the foot arepih-lsiscuffle, make scuffling noises,pih-wyi:sirub with the foot,phe-pweslip in wetor muddy places, andpih-ymitrot.

In her reconstruction of Proto-Yukian (from the three Yukian dialects or closely related

languages), Schlichter (1985: 43) notes that a few derivational, non-pronominal prefixes are

reconstructable. She lists *mi- of the hand or foot, *n- of the mouth, *hu- face, and a fewothers of less clear meaning. As in Wappo, the prefixes appear in both nouns and verbs. Sawyer and

Schlichter (1984) provide several examples of Yuki prefixed verbs, including me-tiik squeeze,

mi-htstep, na-tam-taste, eat, na-koh-teach, and na-hah-drunk, crazy.

3.4 Uto-Aztecan

The Uto-Aztecan family is large in terms of both number of languages and geographical spread. It

contains over 30 languages and extends from Idaho in the north to El Salvador in the south, and

from the California westward as far as Oklahoma, due in part to recent migration. Means/manner

prefixes occur in just one branch of the family: Numic. Numic languages are spoken throughout the

Great Basin, in eastern California, adjacent Oregon, Idaho, and into Utah and Wyoming.

DescribingKawaiisu, a Southern Numic language spoken in south-central California to the east

of the Chumash languages, Zigmond, Booth & Munro (1990: 78) report:

Most typically, these prefixes specify the instrument or body part with which an action was

performed, but they sometimes are used to refer to an object rather than an instrument, or they

can have a less specific adverbial meaning. There are six of these prefixes in Kawaiisu:

ca- with the hand; graspingi- with a long pointed instrument

k- with the mouth or teeth

ma- with the handta- with the foot

w- with an instrument.

Examples of verbs with ca- with hand, grasping are ca-ga- to lead into; to make a twined

basked, ca-nikwi-to pull, ca-pugwii-to sew, mend, ca-waga-to scratch, with intent to drawblood, ca-wee-to scratch an itch, and ca-wrugwi-to shake hands.

Other Numic languages show similar inventories of means/manner prefixes. Tmpisa

(Panamint) Shoshone, a Central Numic language, was spoken before contact in what is now

southeastern California and southwestern Nevada. Dayley (1989: 93-97) lists the following prefixes


11/24

154 Marianne Mithun


which typically indicate the instrument with which a transitive activity is done, but sometimes they

indicate the manner of the activity, and sometimes, especially with intransitive verbs, they indicate

the source or causal factor of an activity (1989: 92-3). ma-/mo- with the hand, k-/ku- with the

teeth or mouth, ku-/ko- with heat or fire, mu-/mo-with the nose, ni- with words, by talking,pa-/po- with or pertaining to water, pi- with the butt or behind, s- with or from cold, sun-with

the mind, by feelings or sensing, ta-/to- with the foot, ta-/to- with a hard rock-like instrument,

to- with the fist, by violent motion, tsa-/tso- by grasping, in the hand, tsi- with a sharp or

pointed instrument, tso- with the head, and w-/wo- with an elongated instrument, generic.

(Many automatically geminate a following consonant.) The meanings added by the prefixes can be

seen by comparing sets of verbs based on the same root, like kaahbreak (of flexible object): ku-

kkaah break from heat, k-kkaahbite in two, break with the teeth, mu-kaahbreak with the

nose,pi-kkaahbreak with the butt, skkaahbreak from cold, ta-kkaahchop with a rock-like

instrument, tsa-kkaahbreak by pulling apart, tsi-kkaahcut, w-kkaahchop.

The inventories of means/manner prefixes in the various California languages vary considerably

in elaboration, but prefixes with certain types of meanings appear in set after set, meanings such as

with the hand or by grasping, with the foot or by kicking, with the teeth or by biting, with fire

or by burning, with a pointed instrument or by thrusting or poking.

4. Central Pomo locative/directional suffixes

The second part of the shared verbal structure consists of sets of suffixes indicating location and/or

direction. Central Pomo, seen earlier, contains such a set. Their effect can be seen by comparing the

verbs in (3), all based on the root -, used for one person running or one or more riding in a

vehicle. (The forms below do not specify tense, but like all Central Pomo verbs, they are inflected

for aspect. Perfectives are marked by -w after a vowel and zero after a consonant. Imperfective

singulars are marked by -n.)

(3) Central Pomo suffixes

-w run (one)

-:la-w run down

-:qa run up (as up a hill)

- run away

-way run against hither, as when a whirlwind came up to you

-:w-an run around here and there

-mli-w run around it (a tree, rock, house, pole etc.)

-ma run northward

-:q run by, over (along on the level), southward

-m run over, on, across (as bridge)

These locative/directional suffixes can and often do co-occur with the means/manner prefixesdescribed earlier.

(4) Central Pomo prefix-suffix combinations

da-d-:la-w push something over a cliff

ma-d-:la-w kick something over a cliff

pha-d-:la-w slowly glide into a swimming pool

ph-d-:la-w jump down, over a cliff, into the water

a-d-:la-w chase (dog) downhill

ba-d-:la-w walk downhill singing

-d-:la-w carry something downhill in hands

-d-:la-w carry something downhill by the handle


12/24



Like the means/manner prefixes, the locative/directional suffixes are pervasive in the vocabulary

and in speech.

5. Location and direction beyond Pomoan

Cognates of the Central Pomo locative/directional suffixes can be found in all languages of the

Pomoan family, so they can be reconstructed for Proto-Pomoan (Oswalt 1976: 23-24). Suffixes of

this type occur beyond the family borders as well. As we will see below, the general structures are

the same, though the shapes of the markers are not.

5.1 Location and direction within the Hokan group

Inventories of locative/directional markers are quite elaborate in some of the languages that have

been proposed to be members of the Hokan stock.

Karuk, the Northern California isolate, contains a rich set, many of which reflect the setting

along the Klamath River in which the language is spoken. Bright (1957: 95) lists the following: -mu

thither, -rupu hence downriverward, -unih down from a considerable height, hence

downhillward, -ra: hither, hither from downriver, hither from downhill, -ura: up to a

considerable height, hence uphillward, -faku hither from uphill, -ro:vu hence upriverward, -

vrak hither from upriver, -sip(riv) up to the height of a man or less, -i(rih) down from the

height of a man or less, -kahence across a body of water, -rinahither from across a body ofwater, -kara horizontally away from the center of a body of water, into ones mouth, -rPa:

horizontally towards the center of a body of water, -rPa:out of ones mouth, -rmnihinto a

container, rukout of a container, -varain through a tubular space, -kivout through a tubular

space, -rprihin through a solid, -rpravout through a solid, -frukinto an enclosed space, -rPukout of an enclosed space, -vrinin opposite directions, -tunvatoward each other, vrayva

here and there within an enclosed space, and -una here and there in an open area. Someexamples of the use of -kara horizontally toward the center of a body of water are ikpuhswim,

ikp:h-kara swim across the river; ikxip fly, ikx:-kara fly across the river; ikakjump, ik:-kara jump into (a river or lake); piv- step, p:v-kara step out over (a creek); sir disappear,s:n-karaswallow; trivpour, and triv-karadrink.

Shasta, spoken immediately to the east of Karuk in Northern California and into Oregon, also

shows a set of locative/directional suffixes. Silver (1966: 144-155) lists -cw away, thither, -eh:hawacross, -eh:wdownstream, -eh:wout of a container, -aha:binto a circumscribedarea, -wa:k in liquid, -haNpi- downward, outward, -asw away, off, -cw upward, -iNbthrough a circular or tubular space, -ipsirudown into, as well as some suffixes which occur in

combination with others, -waklocated in relation to a circumscribed area, -uhialong with, after, -

h:i into, -tac: toward, -kn up over, -kway up along, -:i down along, -awhi straightahead into a circular/tubular space, -wa:kin liquid, -ak:encircling a long object, -ah:walongin space, -ka:h upstream from the mouth of a stream, and -rakmaki here and there. Some

examples of the combinations are: kwruhikwaya:ka he walked along the river upstream,r:eawhi Let him look straight ahead into the hole, kwhuhi:hmpik he puts mouth on bodyand sucks, and kxam:ehempirakmak:ira:hes going from chair to chair (to see which is the mostcomfortable).

The Palaihnihan languages also contain locative/directional suffixes. Talmy (1972: 432-467)

lists numerous examples of Atsugewi verbs containing them. Among them are -ak:locative, -ik:

hither, -imthither, -icup, -micdown onto the surface of the ground, -uwwoff from over a

surface, -ict into a liquid, -iks at the lateral surface of a solid, -cisw down onto the upper

surface of a solid, -ihiy on ones body, -cam to a position athwart a firesite, -asw all about

within itself (hair tousling about, clothes flapping about), and -a:sy multiply together, into an

accumulation. The suffixes also occur in combination:-hiy-ik:out of a snug enclosure or socket,


13/24

156 Marianne Mithun


detached from moorings, -nik-iyall about, here and there, back and forth, -pik-aywaround, -m-

ik:onto a head, into a face, into an eye, -wam-iminto someones body, -tip-aswapart, and -

tip-u:down into a pit in the ground. This last can be seen in nohmqthpu:mathe house fell all the

way down into the cellar (nw-uh-miq-tipu:-im-a EVIDENTIAL-by.gravity-for.house-like.structure.to.move-down.into.pit-thither-FACTUAL). Another example is sphop:oqw:eh: I

blew the dust off my clothes (sw-phu-puq-uww-ihiy-aFACTUAL-blowing-dust.to.move.into. cloud-

off.from.over.a.surface-.on.ones.body-FACTUAL)

Yana contains a similar set of markers. In his introduction to the Yana dictionary, Swadesh

remarks that the most numerous class [of suffixes] is directionals and locations (Sapir & Swadesh

1960: 13). Sapir lists the following suffixes for the Northern Yana dialect: -ai-in the fire, -at-

up to, arriving at, -bal-up from the ground, -bil-about, here and there, -di-bil-at random, all

over, to pieces, -bit-following, -bu- first, ahead, -dap- out of the house, -di- down, -du-

back home, retracing ones steps, -du-straight down, moving down, -duni- one after another, -

du-nitdi- back, backwards, -da-straight ahead, -djil-around, in a circle, -djai:ri-on top, on a

surface, -gap-from the north, -gil-moving over the crest, -git-to a neighboring house, -hau-tothe east, -ki-toward the speaker, hither, -kal-sa-moving through, -kap-to oneself, toward the

speaker, -kul-sideways, slantwise, -lau-out of an enclosed space, out of the mouth, out of ones

eye, from a spring, from the plain up to the mountain, out of the ground, from the water on to land,

-lu- in the head, -ma- there, from there, -mari- down in a hole, pit, basket, -malli- just

outside the house, -mhdji-to the west, -m

hgu-near the house, right there, -nai-sa-all over the

house, -sa-off, away, -sdja-going up (refers to houses), -rtdja-up in the air,-t

hk

hi-from the

east, -thp

ha-to the south, -

hthau-in the plain, -gai-all over (hands and feet), -tstsirat-in a

zigzag line, -tsgil-in the water, to the water, -uldi-down from a height, down the mountain,

-wadju-from the south, wagal-through a hole or opening, -wasa-above, over, -wal-di-down

on the ground, walmi- in hiding, in the woods, -walsa- hidden in the house, -wil- across (a

stream), crosswise, -wil-mi-on one side, -wul-into the house, into an enclosed space, and uimai-

in the middle, between (Sapir 1922b: 224-230). A number of verb roots never appear without suchsuffixes, particularly those describing position or movement, such as one male walks, dance,

chase, extend ones foot, flow, spill, turn, spread out, reach out, handle, long object

moves, blow, suck, creep, shoot, wind to blow, and more. The root i- creep, move in

non-erect fashion, for example, appears in verbs iaisi he lies down in fire (Yahi dialect),icgil- to get into a spring to bathe, ihduwunsiti (squirrel) scurries back into its burrow

(Yahi), ikudimasi he slips up from behind, and ikudsasi he slinks away, creeps away(1960: 199).

Washo also contains a rich set of locative/directional markers including -mad to the south,

from the north, -mii to the east, from the west, -tidown, downwards, -li-weup, upwards, -:ti-we uphill, diagonally upwards, -we on the ground, -ps up from a surface, -:dabspread over a horizontal surface, aon something,i-lmunder, underneath, -had across, -wd over the edge, the summit, -:in, -tum-uwedown in, out of sight, -madin a tube, -mi out from, -yab through a narrow opening, going past, -:gal in the mouth, down thethroat, -we in a certain direction, :ib straight, right, correct, -mkum arched, -:kl hereand there, back and forth, -:gel around something, around in a circle, -yu oscillating, -llb

together, -t:tmapart, andi-y away, out of the way (Jacobsen 1980).

5.2 Location and direction in the Penutian group

Like the means/manner prefixes, the locative/directional suffixes occur beyond the Hokan group. In

fact they are as pervasive among Penutian languages as among Hokan languages, if not more.

The Maidun languages all show sets of locative/directional suffixes. Maidu contains 15 such

suffixes, all but one of which (-ta on top of) imply motion. Shipley (1964: 42-44) lists the


14/24


15/24

158 Marianne Mithun


more preliminary Hokan and Penutian stocks. Yet we see striking parallelisms in abstract

grammatical structure which cross-cut genetic lines. The languages are characterized by pervasive,

often elaborate sets of means/manner prefixes and locative/directional suffixes, structures that are

relatively rare outside of the area. The situation is strongly suggestive of transfer through languagecontact. Relatively little is known for certain of the prehistory of the California groups before their

contact with Europeans, but it is clear that there was an extensive period of intense social contact,

multilingualism, and intermarriage in this area. Most of the communities have always been small.

Because of the richness of the terrain in food sources, vast territories were unnecessary for

subsistence, and exogamy was common.

But it is not immediately obvious how such fundamental structure could be borrowed. The

shared structures are deeply embedded in the grammar. We know that bilingual speakers easily

transfer vocabulary and even phrases from one language to another. If bilingualism is sufficiently

widespread, the adoption of vocabulary can bring with it the adoption of new sounds as well, as

borrowed words are rendered with their native pronunciation. We have seen such a process, for

example, in the adoption of the voiced fricative of French rougeand garageinto English. Heavyborrowing of vocabulary can bring about the transfer of specific affixes as well. If enough sets of

morphologically related words are borrowed, speakers of the receiving language may perceive

morphological patterns among them and extend affixes they contain to native vocabulary. We can

see the adoption of numerous affixes from French into English by this route, such as the repetitive

prefix re- from pairs such as view/re-view and place/re-place. This prefix has been extended to

verbs not of French origin, such as re-do, re-shelve, and many more. Such a process does not appear

to underlie the structural similarities among the California languages, however. The markers

themselves are generally cognate within families, but the forms do not match across genetic lines.

How could speakers transfer abstract patterns such as means/manner prefixes and

locative/directional suffixes without transferring the affixes themselves?

It may not have been the modern morphological structures that were borrowed as such, but the

precursors to them, the patterns of use from which they subsequently developed. We know thatbilinguals accustomed to expressing certain semantic distinctions in one language often transfer

those patterns of expression into another. If, for example, members of one speech community

specified manner of motion or direction particularly often, it would not be surprising if speakers of

that language tended to specify these features often when speaking another. If a particular

grammatical structure existed in both languages, and it was exploited particularly heavily in one

language, it would not be surprising if bilinguals transferred this frequency of use to the other

language.

6.1 Compounding

The languages themselves provide evidence of a probable pathway of development from frequent

patterns of expression to the grammatical structures we see today. The means/manner and

locative/directional constructions seen throughout California and adjacent areas appear to bedescended from compounds. We know that compounding is an extremely common word formation

process cross-linguistically, and compounds can be found in most of the modern languages under

discussion here. Central Pomo contains such NOUN-VERB VERB compounds as kh manw

dance (dance foot-set), ma qaw:n eat (food biting-go), and sm qhn: snore (sleepgrowl). Though both elements of each compound exist in the language as separate words, they

behave phonologically as single words in connected speech in terms of vowel harmony, alternating

length, and stress. The verb kh manwdance, for example, is usually pronounced k

h m:nw.

Similar compound constructions can be seen in Yuman languages. Miller (1990: 50) lists NOUN-

VERB VERBcompounds in Jamul Diegueo such as mii mekuupayfollow (miileg, foot), xa

nup bathe (xawater), yay xanbe happy (heart-be.good), yay mewalybe soft-hearted, and


16/24



yaymat shemabe crazy (heart.and.body-sleep). Most of the stems in these compounds still occur

independently. Describing compounding in the closely related Mesa Grande Diegueo, Langdon

notes that Compounding is usually accompanied by a certain amount of reduction of unstressed

vowels (shortening, omission, or reduction to //). Some consonants are also occasionally omitted.(1970: 133-134).

Describing Maidu, Shipley (1964: 38-40) provides such NOUN-VERB VERB compounds as

pnpe-to smoke (pn-ptobacco-eat), kwb-dig a hole (kw-b-dirt-dig), and cwi-tobe naked (c-wi-clothing-lack), as well as VERB-VERBVERBcompounds tkce-to believe(tk-cehave enough-see), and slid-to sing along with (sl-idsing-perform.together). Thecomponents of these compounds also occur independently as nouns and verbs. In their Yuki

dictionary, Sawyer & Schlichter (1984) list such NOUN-VERB VERB compounds as hawa y yu-cook (food make) and VERB-VERB VERB compounds as koola:l- pace (ko-la:l- walk-ride). Nez Perce shows such NOUN-VERBVERB compounds as qill:wawyacaI sing a departing

song for warriors (qillu:-wwye rawhide-beat) and VERB-VERB VERB compounds as

inupcikl:toqima I came back to get it (np-cikl seize-turn) (Aoki 1970: 88-89). If certainpatterns of compounding were particularly productive in one language, it would be easy forbilinguals to transfer the propensity to form compounds on that model into another language.

In fact we can see exactly the kinds of NOUN-VERBand VERB-VERBVERBcompounds that

would give rise to the means/manner prefixes in the area. Compounding is pervasive among all

branches of the Uto-Aztecan family, including Numic. The initial noun roots in NOUN-VERB

VERBcompounds, a type also called noun incorporation, often designate the means or instruments

by which actions are carried out. For Kawaiisu, for example, Zigmond, Booth, and Munro (1990)

cite such compounds as moo-paa- stir by hand (hand-stir). The noun root moo also occursindependently. VERB-VERB VERBcompounding is also common, as in kaa-pagi-to walk andeat (eat-walk). For Tmpisa, Dayley (1989: 91-92) lists such compounds as kikuttihjab with the

elbow (elbow-shoot) and mitkkaannangkawih speak (in) English (whiteman-speak). In his

Tmpisa grammar, Dayley traces the sources of nearly all the Tmpisa means/manner prefixes toProto-Uto-Aztecan noun and verb roots.

(5) Lexical sources of Tmpisa prefixes: Dayley (1989: 93-97).

Tmpisa Proto-Uto-Aztecan

ma-/mo- with the hand *maa hand

k-/ku- with the teeth or mouth *ki bite

ku-/ko- with heat or fire *kuh fire

mu-/mo- with the nose *mu-pi nose

ni- with words, by talking *niya/niha name

pa-/po- with or pertaining to water *paa water

pi- with the butt or behind *pih- back

s- with or from cold *sp coldsun- with the mind, by feelings or sensing *sunna heart,

*suuwah believe

ta-/to- with the foot *tannah foot

ta-/to- with a hard, rock-like instrument *tn rock

tsa-/tso- by grasping in the hand *tsai grasp,

hold (Numic)

tso- with the head *cohngi head

w-/wo- with an elongated instrument, generic *wpaa whip

Most of these reconstructed Proto-Uto-Aztecan noun and verb roots still survive as roots in the

modern Numic languages, including Tmpisa mooi and Kawaiisu moo hand, Tmpisa: kunnafire(wood) and Kawaiisu kuna fire, Tmpisa mupin and Kawaiisu muv- nose, Tmpisa niya


17/24

160 Marianne Mithun


and Kawaiisu niyaaname, Tmpisapaaand Kawaiisupoowater, Tmpisa suwaand Kawaiisukuu-sb want, feel, Tmpisa tmpe and Kawaiisu tbi rock, Tmpisa tsai/tse grasp, hold,Tmpisa tsoppipphhead, and Tmpisa wa, Kawaiisu waapenis or Kawaiisu kwipawhip.

We can also see examples of the kind of compounding that could be the precursor to the

locative/directional suffix constructions. Some come from languages of the Chinookan family,

spoken along the Columbia River on the northern Oregon border. Hymes describes a pattern of

VERB-VERBVERBcompounding in Kathlamet Chinook. The small set of roots that can appear in

second position all indicate motion in a particular direction. They include -pamotion out of, -pck

motion from open to cover, especially from water to shore or inland, -pq motion into, -tk

motion, position down, -Lx motion from cover to open, especially toward water, -cu motion

down, and -uulXmotion up. Some examples are ikai-cuit fell (-i-cugo-down), ikuXni-pck

she drifted ashore (-Xuni-pckdrift-from.water.to.shore), iskiski-Lx they two launch it (-ski-Lx

launch-motion.toward.water), and iciskna-pq he looked into (house) (-na-pq look-into)

(Hymes 1955: 218). These verb roots also appear independently. The root -pa to go outcan be

seen in anu-p-iaI will go out. The root-pckmotion from open to cover, especially from waterto shore is in tx-pck-aLet us go inland and ikppck-ashe went up. The root -pqmotion into

appears alone as enter in alx-pq-a we will go in. The root -tk motion, position down is

translated put down, place on its own: icLX-tkhe put it on the ground. The root -Lxmotion from

cover to open, especially toward water, can be seen in a-Lx-tshe comes down to the water.

For the most part, the root sources of the modern means/manner prefixes and locative/directional

suffixes no longer survive in the other languages, but in a few cases, traces remain. Earlier we saw

root sources of the Numic means/manner prefixes in Kawaiisu and Tmpisa. Jacobsen comments on

possible origins of the Washo prefixes (1980: 86).

Only a few rather tenuous resemblances have been found between four lexical prefixes and

semantically related stems or nominal prefixes.. The prefux dul- with the hand, descriptive of

hand resembles the noun stem :du hand and also the prefix tul- occurring in tul cig

/tulcik/ finger, toe and tul pi nail, claw. The prefixes mege- to be washed along andm

ed-, m

ed-to float resemble the noun and verb stem mewater, to drink. Less compellingis the resemblance of the prefix cig- with the buttocks, descriptive of buttocks to the noun

stem cabuttocks.

Describing the Wappo prefix hu- head, Radin (1929: 32) writes, It is always an old

incorporated noun that is felt here. Barbareo Chumash contains a manner prefix ti-speaking,

seen in such verbs as ti-molotell stories of olden times (moloqlong ago), ti-yepteach, show, ti-

kuytalk about, and ti-patuninterrupt, butt in on, and ti-pawilto talk. It is apparently descended

from the noun tiname (Beeler 1976). Some of the languages, however, show us much more.

Sources of some locative/directional suffixes can still be seen in the Pomoan family, even within

Central Pomo itself. Among the Central Pomo directional suffixes is -manorthward, visible in

such verbs as -ma (one) ran northward, h-ma (group) ran northward, m-ma (oneelderly person) walked, (one) crawled up or northward, and p

ha-d-ma(one) swam upstream.

The source of this suffix still survives as an independent verb m go north. Another of the

Central Pomo suffixes is -m(a)over, across, on. It can be seen in -m(one) ran over, on, across

(as bridge), h-mgroup ran over (bridge), -d-mled group across, and ph-d-m(one) jumped

across. The source of this suffix survives as a verb root in the language as well: -mgo over on

foot, walk on or across. Inflected with the perfective aspect suffix -w it appears in m-w(s/he)

went over on foot, walked across. The same root appears with prefixes in the usual verbs used for

sitting on something like a rock, log, or chair, as opposed to the ground: h-m-w(one) sits, ba-

m-w (group) sits. The -m suffix has also developed a slightly more abstract use with verbs of

position to indicate a constant state: d-mone is holding it (such as an empty cup) there, ph-d-m


18/24



(it) (like a flag) is just hanging there, -d-mit (something with a handle) is hanging there, one is

holding it by the handle.

6.2 Bipartite stems

In some families in the area, the means/manner prefixes and locative/directional suffixes appear to

be ancient, already in place in the common parent language. All of the prefixes and suffixes in the

modern Pomoan languages can be reconstructed as such for Proto-Pomoan. The Yuman prefixes

appear to be quite old as well and are now often difficult to identify. The small set of means/manner

prefixes found in Wappo and Yuki can be reconstructed as prefixes for their common parent. But a

number of languages in Northern California and Oregon show an intriguing intermediate stage of

development.

In his description of Northern Yana, Sapir characterized the initial means/manner morphemes

like d- do with the fingers, bui- do with the feet, dja- eating, have in the mouth, and dju:-

wind blows as stems, but he specified that they could not be inflected directly with tense, mode,

and pronominal suffixes. He noted that each of these morphemes, which he distinguished by

asterisks in his stem list, must have its meaning completed by one or more derivational suffixes

(1922a: 217). In the long list entitled non-grammatical verb suffixes are such morphemes as -

balla-to push, knock, -di-to clean, -di-to come loose, gidi-fire goes out, -tsut-to split,

burst through, and many more that would be classified as roots in other languages. It seems at first

surprising that the initial elements were not categorized as prefixes and the second elements as

roots. The rationale behind his decision to categorize morphemes like d-with the fingers and bui-

with the feet as stems is not made explicit, but it could be due in part to their concrete meanings

and relatively large numbers. Furthermore, his list of non-grammatical verb suffixes also includes

locative/directional markers such as -bil about, here and there, -di- down, -du- back home,

-djil-around, in a circle, -djai-ri-on top, on a surface, -gap-from the north, and others seen

above.

Similarly, the Maidumeans/manner morphemes such as ka-with the flat hand or flat side ofsomething that Dixon listed as prefixes are classified as stems by Shipley in his dictionary (1963).

Like Sapir, Shipley notes specifically that these morphemes never occur as stems on their own. A

sample entry is h-: a morpheme having to do with thinking and feeling in general. This form does

not occur alone as a stem. The examples cited of its use bear this out: hbokjnot know how to,

hdan be preoccupied with, hhje think, hkes wise, hkicik forget, hkit lonesome,

hkinuremember, hmitrecall, bring to mind hnorely on, hpajguess, and hwjethink

(Shipley 1963: 120).

In the same way for Klamath, Barker classifies means/manner morphemes such as l-acting

with the fingers, fingernails and v-acting with a long object as stems, with the proviso that theydo not occur directly before a modal morpheme, that is, they cannot be inflected directly (1964:

115). In his grammar of Sahaptin, Jacobs terms means/manner morphemes such as pa-with the

hand or fingers anterior roots, noting, In listing anterior, central and posterior stems it will be

assumed unless otherwise indicated that each element given appears only when compounded with

some other independent or similarly non-independent stem (1931: 154).

In a similar way, Jacobsen classifies the locative/directional morphemes of Washo such as -

mad to the south, from the north, -ti down, downwards, etc. as stems, with the provisothat they cannot occur on their own. Jacobs classifies Sahaptinlocative/directional morphemes

such as -wi-down and -tuni-upstream as posterior stems.

It seems at first strange that these authors did not simply classify all of the means/manner

markers as prefixes, and the locative/directional markers as suffixes, if they never constitute a full

verb stem on their own but must always be followed or preceded by another morpheme. The story is


19/24

162 Marianne Mithun


actually more intriguing, however. Consider the Maidu means/manner wi-with the hands, which

Shipley notes does not appear on its own.

(6) Maidu wi-with hands: Shipley (1963: 187-189).

wi ..perform an action with the hands. This morpheme occurs only as first member of compounds.

wi-dk grasp with the hand wi-tl roll (something) up in a

roll

wi-bk detach from a flat surface wi-cp dislodge

wi-cmyl stir a small quantity (e.g. coffee) wi-dsdo rip, split

wi-hp pull (long, stemlike objects) wi-tbil twist, wind on

wi-hl pull (anything) wi-hp stretch by pulling

wi-k tug at wi-kp extract

wi-kj lift wi-kl perforate

wi-pl pluck, pull out, extract wi-pn tie

wi-sp break by snapping suddenly wi-sp snatch

wi-tm cinch, tighten etc.

Verbs containing wi-can also contain a locative or directional suffix.

(7) Maidu locative/directional suffixes: Shipley (1963: 187).

wi-dk-doj grasp with the hand and pull up -doj upward

wi-dk-daw pull off, pull loose with the hand -daw down and away

wi-dkha-kj lead someone away by the hand -kj away thither

wi-dk-no-je lead someone around by the hand -no-je along-hither

wi-dk-paj break off with the hand -paj against

What is surprising is that Maidu stems can be made up of just a means/marker and a

locative/directional marker.

(8) Maidu means/manner + locative/directional: Shipley (1963: 187).

wi-dj pull (something) up pulling-upward

wi-wj divide pulling-apart

If we were to analyze all of the means/manner markers as prefixes, and all of the

locative/directional markers as suffixes, we would have verbs without roots.

The same phenomenon occurs in Yana, Klamath, and Sahaptin. These languages are part of what

DeLancey (1996), following Jacobsen, has termed the Bipartite Stem Belt. The languages contain

large numbers of verb stems consisting of at least two parts. Those markers that specify means or

manner or characterize the entity involved appear initially and must be followed by another stem

element. Those that specify location or direction must occur finally and must be preceded byanother stem element. The bipartite stems may thus differ from prototypical compounds in being

composed of elements neither of which can occur as a root on its own. In his 1980 description of the

structure in Washo, Jacobsen described these dependencies.

We are considering a set of over 125 prefixes of concrete meaning, which occur with members

of a set of around 200 stems, also having concrete meanings ... With minor exceptions, the

members of the one class occur only with members of the other, thus making a closed system.

Since the verb stems do not occur independently, we may call them dependent verb stems.... In

so far as a system like this is completely closed, it would be difficult or meaningless to

differentiate between stem and affix. This is not quite the case, however, and at least two

considerations point to the second member as being the stem.


20/24


21/24

164 Marianne Mithun


languages, appear to be older than others, such as those in the Numic languages: in the Yuman

systems, the lexical items that served as the sources of the prefixes can no longer be found, and the

meanings of individual markers can be difficult to identify, while in the Numic systems, lexical

sources of nearly all of the prefixes still survive as independent roots, with nearly the same shapesas the prefixes. Numic speakers, situated at the periphery of the linguistic area, could have

accelerated the reduction and solidification of their compounding constructions into the closed

prefix system under the influence of their neighbors.

7. Conclusion

If we are to tackle the issue of the relative susceptibility to borrowing of grammatical structures, it

is important to add the dimension of time to the investigation. Many abstract structures that might,

on the surface, seem impossible to borrow, can nevertheless be the result of contact. It may not have

been the modern structures that were borrowed, but rather the precursors to their development.

As we have seen, the California linguistic area is characterized on the one hand by tremendous

genetic diversity, and on the other by fundamental, deeply-entrenched morphological structures that

cut across genetic lines, structures that are comparatively rare outside of the area. A striking

proportion of the languages exhibit a verb structure characterized by often elaborate systems of

means/manner prefixes and locative/directional suffixes. The semantic and structural parallelisms

observable across genetic boundaries, along with the longstanding history of multilingualism and

intermarriage in the area, suggest that the similarities are the result of contact. But it is difficult to

imagine how such abstract systems could be borrowed, particularly without the individual

morphemes that carry it.

In fact the prefix and suffix constructions need not have been transferred directly. Evidence

suggests that what was transferred were patterns of expression, tendencies to specify particular

distinctions especially often and to exploit existing constructions, compounds, for these purposes.

The shared affix patterns described here appear to have developed from the compound constructions

by a relatively straightforward process, the crystallization and reduction of frequently-occurringinitial roots specifying means and manner, and final roots specifying location and direction. The

spread of the original patterns of expression did not necessarily occur simultaneously over the area.

The subsequent development of roots into affixes may have taken place independently within the

individual languages, or it may have been stimulated by further contact. Such reduction of roots into

affixes is not the only way in which grammatical systems may develop in parallel from shared

patterns of expression. When frequencies of use are transferred from one language into another, the

scene can be set for a wide variety of reanalyses.

An appreciation of the diachronic dimension in studies of language contact permits to enlarge

our view of the kinds of grammatical developments that can be attributed to contact, which can in

turn be important in judging evidence for deep genetic relationships. Typological similarity has

been taken as evidence of common origin. In fact the Hokan and Penutian hypotheses were initially

based primarily on typological similarity rather than recurring sound correspondences. (It should be

noted that the prefix and suffix structures described here were not considered part of the evidence.)

The use of typological evidence depends on the widely-held idea that grammatical borrowing can

take place only between languages that are very similar typologically. Such an idea certainly seems

reasonable. It should be impossible to transfer a structure into a language where it will not fit. But

typology can shift over time, step by step. The shift may begin with features that are relatively easy

to borrow, such as a change in relative frequencies of alternative word orders. Such shifts can in

turn, given sufficient time, lead to the development of similar orders of morphological categories

within words, a typological feature it would be difficult to borrow. The possibility of gradual

typological alignment indicates that structural parallelism alone cannot provide conclusive evidence

of deep genetic relationship.


22/24



The more we can learn about the ways in which grammatical structure can develop over time,

the more we will learn about the potential impacts of language contact, and vice-versa.

References

Aoki, Haruo. 1970. Nez Perce grammar. University of California Publications in Linguistics 62.

Berkeley: University of California.

Beeler, Madison. 1976. Barbareo Chumash Grammar: A Farrago. Hokan Studiesed. by Margaret

Langdon & Hirley Silver, 251-269. The Hague: Mouton.

Bright, William. 1957. The Karok Language. University of California Publications in Linguistics

13. Berkeley: University of California.

Campbell, Lyle. 1997.American Indian Languages. Cambridge: Cambridge University Press.

Dayley, Jon. 1989. Tmpisa (Panamint) Shoshone Grammar. University of California Publications

in Linguistics 115. Berkeley: University of California.

DeLancey, Scott. 1988. Klamath Stem Structure in Genetic and Areal Perspective. Papers from the

1988 Hokan-Penutian Languages Workshop. University of Oregon Papers in Linguistics 1:

31-39.

DeLancey, Scott. 1991. Chronological Strata of Suffix Classes in the Klamath Verb. International

Journal of American Linguistics57: 426-445.

DeLancey, Scott. 1996. Penutian in the Bipartite Stem Belt: Disentangling Areal and Genetic

Correspondences.Berkeley Linguistics Society 22: Special Session on Historical Topics in

native American Languages. 37-54.

DeLancey, Scott & Victor. 1997. The Penutian Hypothesis: Retrospect and Prospect. International

Journal of American Linguistics63: 171-201.

Dixon, Roland. 1911. Maidu. Handbook of American Indian Languages. Bureau of American

Ethnology Bulletin40.1: 670-734.

Dixon, Roland & Alfred L. Kroeber. 1903. Native Languages of California. American

Anthropologist5: 1-26.

Dixon, Roland & Alfred L. Kroeber. 1913a. Relationship of the Indian Languages of California.

Science37: 225.

Dixon, Roland & Alfred L. Kroeber. 1913b. New Linguistic families in California. American

Anthropologist15: 647-55.

Heine, Bernd & Tania Kuteva. 2005. Language contact and grammatical change. Cambridge:

Cambridge University Press.

Hymes, Dell. 1955. The Language of the Kathlamet Chinook. Ph.D. dissertation, Indiana

University.

Jacobs, Melville. 1931. A Sketch of Northern Sahaptin Grammar. Seattle: University of Washington

Publications in Anthropology 4.2: 85-292.

Jacobsen, William. 1980. Washo Bipartite Verb Stems.American Indian and Indoeuropean Studies

ed. by Kathryn Klar, Margaret Langdon & Shirley Silver, 85-100. The Hague: Mouton.


23/24

166 Marianne Mithun


Langdon, Margaret. 1970. A Grammar of Diegueo: The Mesa Grande Dialect. University of

California Publications in Linguistics 66. Berkeley: University of California.

Langdon, Margaret. 1990. Morphosyntax and Problems of Reconstruction in Yuman and Hokan.

Linguistic Change and Reconstruction Methodology ed. by Philip Baldi, 57-72. Berlin:

Mouton de Gruyter.

Miller, Amy. 1990. A Grammar of Jamul Diegueo. Ph.D. dissertation, University of California,

San Diego.

Mithun, Marianne. 1999. The Languages of Native North America. Cambridge, UK: Cambridge

University Press.

Oswalt, Robert. 1976. Comparative Verb Morphology of Pomo, Hokan Studies ed by Margaret

Langdon & Shirley Silver, 13-28. The Hague: Mouton.

Powell, John Wesley. 1891. Indian Linguistic Families of America North of Mexico. Bureau of

American Ethnology-Annual Report7:1-142. Reprinted 1966 in: Lincoln, NB: University ofNebraska Press. Preston Holder, (ed.).

Radin, Paul. 1929. A Grammar of the Wappo Language. University of California Publications in

American Archaeology and Ethnology27: 1-194.

Ross, Malcolm. 2001. Contact-induced change in Oceanic languages in northwest Melanesia. Areal

Diffusion and Genetic Inheritance ed. by Alexandra Aikhenvald & Robert M.W. Dixon,

134-166. Oxford University Press.

Ross, Malcolm. 2006. Metatypy.Elsevier Encyclopedia of Language and Linguistics.

Sapir, Edward. 1922a. The Takelma Language of Southwestern Oregon. Handbook of American

Indian Languages, Franz Boas (ed.).Bureau of American Ethnology Bulletin40.2: 1-296.

Sapir, Edward. 1922b. The fundamental elements of Northern Yana. University of CaliforniaPublications in American Archaeology and Ethnology13: 215-234.

Sapir, Edward & Morris Swadesh. 1960. Yana Dictionary, ed. by Mary Haas. University of


Sawyer, Jesse. 1965. English-Wappo Vocabulary. University of California Publications in

Linguistics 43. Berkeley: University of California.

Sawyer, Jesse & Alice Schlichter. 1984. Yuki Vocabulary. University of California Publications in

Linguistics 101. Berkeley: University of California.

Schlichter, Alice. 2006. Proto-Wintun. University of California Publications in Linguistics 137.

Shipley, William. 1963. Maidu Texts and Dictionary. University of California Publications inLinguistics 33. Berkeley: University of California.

Shipley, William. 1965. Maidu Grammar. University of California Publications in Linguistics 41.

Berkeley: University of California.

Silver, Shirley 1966. The Shasta Language. Ph.D. dissertation, University of California, Berkeley.

Talmy, Leonard. 1972. Semantic Structures in English and Atsugewi. Ph.D. dissertation, University

of California, Berkeley.

Uldall, Hans Jrgen & William Shipley. 1966. Nisenan Texts and Dictionary. University of



24/24



Ultan, Russell. 1967. Konkow. Ph.D. dissertation, University of California, Berkeley.

Zigmond, Maurice, Curtis Booth & Pamela Munro. 1990. Kawaiisu: A Grammar and Dictionary

with Texts. University of California Publications in Linguistics 119. Berkeley: University of

California.

Marianne Mithun_Contact and Time_2007

Documents

Transcript of Marianne Mithun_Contact and Time_2007