Etnopoetica Dell Hymes

Oral Tradition 9/2 (1994): 330-370
Ethnopoetics, Oral-Formulaic Theory,

and Editing Texts
Dell Hymes
John Miles Foley (1992) has opened up a consideration of the

connections between oral-formulaic theory and work that has come to be
called ethnopoetics. This is much to be desired, for until recently the two
have seemed to occupy different worlds, yet a general view of oral poetry
requires both. Foley focuses on a major thrust of folklore today, the
interaction between performance and tradition. Here I want to focus on two
older concerns, the structure of texts and manuscript sources.
Constraints within and among lines
Oral-formulaic theory and ethnopoetics are both concerned with

composition in the course of performance, and with constraints that must be
met in doing so. In the epics and other poetry studied in terms of oral-
formulaic theory, the constraint is a metrical line, commonly a sung metrical
line. In oral narratives the constraint is commonly a relation among lines.1
1
Sung epic poetry is famous for oral formulae, which have been taken as enabling a
narrator to meet the constraint of the metrical line in the midst of performing. (I realize that
there is more to oral formulae than that). The narratives with which I have worked, not
having metrical lines, do not have the same performance constraint. One does sometimes
find evidence of fulfilling the constraint of a patterned sequence in an ad hoc way. Among
Native Americans in the Pacific Northwest, formulae seem to have two roles. Prayers and
exhortations at ceremonies may be full of them, not to meet formal constraint, but to invoke
tradition. Narratives employ them at major junctures, such as openings and closings, and
there are classes of words to be expected as markers. All these could be said to be required
by a genre. There are also words expectable for characteristic actions in the course of a
type of scene or story. The choice, position, and frequency of these words is particular to a
ETHNOPOETICS, ORAL THEORY, EDITING TEXTS 331
When constraint is internal to the line, we do not hesitate to speak of

poetry. In the oral narratives of many Native American peoples and many
speakers of English, perhaps universally, there is a constraint external to the
line. It has to do with relations among lines that count as verses. A
verse is usually easily recognized in speech. It is marked by one of the
main intonational contours of the language. Such verses form sequences,
and do so in terms of a small set of alternatives. There appear to be two
fundamental principles. The usual (unmarked) alternatives may be
sequences of two or four. Many Native American communities (such as
those of the Kwakiutl, Takelma, Zuni, Hopi, and Navajo) make use of such
sequences. Many others (such as those of the lower Columbia and
Willamette Rivers), and many speakers of American English, so far as is
known, connect verses in sequences of three and five.2
Narrators are not restricted to just these alternatives. Some command
both principles, and may adopt one or the other for a particular story or
situation, or part of a story, or level of organization (cf. Hymes 1990-94,
1993b, 1994b).
If organization in lines is a general definition of poetry, then these
narratives are poetry. In one kind of poetry what counts first of all is a
relation within the line, a relation among syllables, stresses, alliterations,
tones, conventional feet. In another kind what counts first of all are relations
among lines (more properly, verses) themselves. If the first kind is metrical,
the second kind can be called measured. It is sometimes called measured
verse, and its analysis, verse analysis.3
Such analysis depends upon three principles. One is that just
discussed. It implies that narratives transcribed and published as prose
paragraphs are in fact organized in lines. The second principle is one
Roman Jakobson considered basic to poetry, and called equivalence (1960).
Sequences however diverse may count as equivalent in the organization of a
given narrator and performance. They seem to give shape as much as to fulfill it.
2
African American narratives collected in New York City by William Labov can
be more accurately appreciated when seen to be poetry in this sense. Labovs much-used
analysis of stories in terms of a set of universal functions misses their shape. The stories
are not a linear sequence of temporal events, intersected by non-temporal effects, but
successive arcs of arousing and realizing expectation. See Hymes 1991, 1994a.
3
There is also of course free verse, much of which actually has recurrences and
relations of various kinds.
332 DELL HYMES
narrative, if some recurrent feature marks them as such. As already noted,

intonation contours are usually such a mark. Sometimes intonation contours
appear not to be such a mark, and verses are signalled by a grammatical
feature, such as the quotative, or a combination of grammatical elements and
patterning itself (cf. Hymes 1982 on Zuni, 1994c on Hopi). For texts for
which we do not know the intonation contours, there still are indications of
equivalence. Sometimes a certain word or words mark the beginning of
units. Turns at talk seem always to count as verses. Other forms of
repetition and parallelism occur.
The principle of equivalence implies a text that is a sequence of units.
In addition to equivalent units (and repetition and parallelism), there is
succession. Succession is not a matter simply of linear sequence, of
counting. Successive units give shape to action.4 In particular, patterns of
succession can be ways of coming to an ending point. As suggested, one
common way is by sequences of two and four, the other by sequences of
three and five.
Sequences of two tend to give to action an implicit rhythm of this,
then that. Pairs of pairs may have the same relation (although other internal
relations may obtain). Sequences of three tend to give an implicit rhythm of
onset, ongoing, outcome. A development of this last, found as far apart as
the Columbia River, Philadelphia, and Finland, integrates two sequences of
three within a sequence of five. It is possible (not necessary) to have the
third unit a pivot, completing one succession of three and beginning
another.5 I call this interlocking. There are other possibilities of rhythm
within each type of sequence, and their representation on the page calls for a
variety of solutions, and a willingness to experiment (see Hymes 1992).
The principle involved in succession became clear to me in rereading
a remarkable essay by Kenneth Burke, Psychology and Form (1968
[1925]). Let me summarize its theme as the recurrent arousal and
satisfying of expectation. Not a straight line, but a series of arcs. What
4
Rhyme and stanza-forms are analogues, especially in a narrative poem. The
difference is that larger units of oral narratives of the sort considered here do not have to
be constant in number of lines or other parts. Narrators need not fill a fixed form.
Rather, they match two sequences as they proceed, one of incident with one of formal
options. The matching can differ from telling to telling. This (re-)generative competence
needs to be taking into account in discussion of entextualization (Bauman and Briggs
1990).
5
Examples are given below.
Burke writes of works accepted as literature is pervasively true of oral

narrative, and often enough deserves the connotations of his other term for
it, eloquence (34-35, 44).
Interlocking in Philadelphia and Yakima (3 and 5). Here are two
examples of interlocking which illustrate the arousal and satisfying of
expectation, and the difference that verse analysis can make. The first is
from Philadelphia. It is from one of a number of narratives collected by
Nessa Wolfson in a study of the use of the historical present. The narrative
has five scenes. Their foci, successively, are a situation [i], the seizing of an
opportunity [ii], acceptance of a bid [iii], acceptance of a demand for
certificates [iv], acceptance of a settlement date [v]. The first three scenes
show an initial condition, development of it, and a proximate outcome. The
series could be taken as complete. The third scene, however, turns out to be
the first of a second series of three, concerned with stages of acceptance.
The two series of scenes interlock at the level of the narrative as a whole.
(The narrative is discussed in Hymes 1993a, but the text itself is not
included).
Interlocking also obtains within two of the scenes ([iii] and [iv]). In
each there are five pairs of verses. The first four are turns at talk between
the realtor and the narrators wife. The fifth relates an outcome, acceptance.
In each series the third pair of verses has the couples offer. It is outcome to
what has preceded, and at the same time an onset to what follows (she wont
accept, she does accept).
Wolfson presented the text in one block paragraph. The lines of the
relevant scenes are shown below between brackets, within part of the
paragraph.
Shes a Widow
...So he says, That you have to do in any house. So she says,

Yes, we have to lay down new floors, the rugs are no good (the rugs
happen to be in good shape), we have totheres too much shrubbery, we
have to tear out some of the shrubs. (The shrubbery around the house is
magnificent if its done right, if its done right.) So really we made up
everything. [So he says to my wife, he says, Well, what would you bid?
So she says, Its stupid for me to talk, she says, You got a bid for thirty-
three, thirty-four, she says. Why should I even talk to you? It aint gonna
be anywheres near. So he says to her, he says, Well, he says, the
person at thirty-four backed out. So she says, Oh yeah? He says,
Yeah, he says, What would you bid? So she says, Twenty-eight. He
334 DELL HYMES
says, Oh, he says, No, that shell never go for. So she says, Okay,
thats my bid, Mr. Smith. You want it, fine; you dont, fine. Got a call
that afternoon. It was accepted! So I go to see the houseI go to sign the
contract, I look at the contract and I says, I aint signing this. He says,
Why? I says, I want a plumbing certificate, I want an air conditioning
certificate, I want a heating certificate, and I want a roof certificate! So he
says, Really, we wont guarantee... I says, I dont want guarantee, I
want certificates, from certified people that its in good shape, and I want
the right to bring in any of my guys. So he says, She wont go for it...
this, that... So I says, Aah, dont be silly, I says, Look, you just take it
to her. So I get a call back about a day later, Okay, shes accepted.] So
then I get anow what I do is, I pick up this thing, I take it to my cousin,
he goes to someone, he says, Settlements no good. Shes got us for forty-
five days. In October she wanted to settle....
Here is how the bracketed passage appears when displayed in terms of

lines, verses, stanzas, and scenes.
[iii] [bid accepted]
So he says to my wife, he says, (A) 60

Well, what would you bid?
So she says,
Its stupid for me to talk, she says,
You got a bid for thirty-three, thirty-four, she says,
Why should I even talk to you? 65
It aint gonna be anywheres near.
So he says to her, he says, (B)

Well, he says,
the person at thirty-four backed out.
So she says, Oh yeah? 70
He says, (C)
Yeah, he says,
What would you bid?
So she says, Twenty-eight.
He says, Oh, he says, (D) 75

No, that shell never go for.
So she says,
Okay, thats my bid, Mr. Smith.
You want it,
fine. 80
You dont,
fine.
Got a call that afternoon. (E)

It was accepted!
[iv] [certificates accepted]
So I go to see the house (A) 85

I go to sign the contract,
I look at the contract
and I says, I aint signing this.
He says, Why? (B)

I says, I want a plumbing certificate. 90
I want an air conditioning certificate,
I want a heating certificate,
and I want a roof certificate.
So he says, Really, we wont guarantee... (C)

I says, I dont want guarantee, 95
I want certificates,
from certified people that its in good shape,
and I want the right to bring in any of my guys.
So he says, She wont go for it... this, that.... (D)

So I says, Aah, dont be silly, I says, 100
Look, you just take it to her.
So I get a call back about a day later, (E)
Okay, shes accepted.
The same relations open a narrative told to Edward Sapir in Wishram

Chinook by Louis Simpson at Yakima, Washington the summer of 1905.
Here are the lines as published in prose paragraphs (Sapir 1909:139ff.)
The Deserted Boy
Some time long ago the (people) said to the boy: Now let us go
for reeds. The boy was (considered) bad. So then they said: Now you
people shall take him along (when you go for) reeds. And then they said
336 DELL HYMES
to them: You shall abandon him there. So then the people all went
across the river. They went on and arrived where the reeds were. And
then they cut off the reeds and said (to them): If the boy says, Are you
people still there? you shall answer him, uu.
And then they all ran off; straight home they ran, went right across
the river. No person at all (was left) on this side; they were all on the
other side. And then that boy said: Now let us all go home!Uuu,
said the reeds to him. He looked about long, but in vain; there was
nobody. And then he too started to go home, he too went following
behind them; he ran until he arrived (at the river), but there were no people
to be seen. So then the boy cried. And then he heard (something)....
Here is the opening in terms of lines, verses, and stanzas:
Now then they told a boy, (A)

Now let us go for reeds.
Long ago the boy was mean.
Now then they said,
Now you will take him for reeds. 5
Now then they told them,
You shall abandon him there.
Now then the people all went across the river, (B)
they went on,
they came to the reeds. 10
Now then they cut them off.
Now then they said,
If the boy should say,
Are you there?,
you shall answer, 15
Uuu.
Now then they ran off, (C)

straight home they ran,
straight across they went,
not a person on this side, 20
all on that side.
Now then the boy, too, said, (D)
Now lets go home.
Uuu,
went the reeds. 25
In vain he searched about:
no person.
Now then he too started home, (E)
he too followed behind them;

he arrived running: 30
now, no people.
Stanzas (ABC) tell of the people deserting the boy. These stanzas are
linked by having the people, they as agents throughout.
Stanzas (CDE) tell of the boy finding himself deserted. These stanzas
are linked by their endings:
not a person on this side, / all on that side.

no person
now, no people
Stanza C is the pivot. The preceding stanzas (A, B) are linked by the
plan to abandon the boy, first by instructions to take him for reeds, then by
instructions to the reeds as to how to deceive and delay him. The following
stanzas (DE) are linked by the boys search for the others. (C) is outcome to
the first pair and the onset for the second. It realizes the plan and provides
the condition for the discovery of absence.
Around (C) indeed there is a chiasmus-like symmetry. The
immediately adjacent stanzas (B, D) involve the instructions to the reeds,
their being given (B) and their being carried out (D).6 The outer stanzas
6
Stanza (D) is a brief form of what can be a full scene. It often occurs in version
of the story-type Bear and Deer. Bear has killed Deer while the two are away from
home. Deers children retaliate by killing Bears children, and flee before Bear returns.
Bear, finding her children dead, pursues them, but first asks a dog the direction they have
gone. The dog has been instructed to bark in turn in directions other than the one in which
the children actually go. Sapir did not record the myth from Louis Simpson (Sapir 1909),
but Victoria Howard dictated it in Clackamas to Melville Jacobs (the incident is in M.
Jacobs 1958:149-50), with both women bears, Grizzly and Black Bear. Charles Cultee told
it to Franz Boas in Kathlamet (the incident is in Boas 1901:122), with neither woman a
bear. I suspect that the doubling in Mrs. Howards version, and the diminution in that from
Mr. Cultee (to Robin [Thrush] and Salmonberry) reflects tension about the figure of a bear
as a way of exploring the nature of women.
In Louis Simpsons The Deserted Boy presumably the reeds answer, first from
one direction, then from another, so that the boy searches everywhere but in the direction
the people have actually gone, to the river. We are to understand that they have taken the
only canoe. The boy is left on a low marshy bit of land (where reeds would grow), too far
from either side of the river for him to get back. Mr. Simpson assumes an audience would
understand this, and subordinates explanation, or elaboration, to severity of form, in which
338 DELL HYMES
have to do with the state of abandonment, its initiation by the people (A) and
its realization by the boy (E).7
Interlocking in Alaska and Cochiti (3 and 4). Native American
narratives taken down in English can display native form. In the summer of
1924 Ruth Benedict took down a number of tales from interpreters from the
Cochiti Pueblo in New Mexico. In pursuit of a type of story involving
Coyote, birds, water, and songs and names imperfectly mastered, I analyzed
one titled by Benedict Coyote imitates Crow (Benedict 1931:149; cf.
Hymes 1994c). The sequence in terms of actions, verses, and scenes seems
clear, probably because it was carefully translated.8 The story has one of the
two examples known to me of interlocking within four-part relations.
For a five-part sequence to contain two interlocking sequences of
three seems possible wherever three and five-part relations are used. Until
early 1993 I knew of but one example with two and four-part relations.
Early in this century the missionary John W. Chapman recorded some
sixteen narratives in the language of those he served (the language has since
been referred to as Ingalik, and now, Deg Hitana). The texts have
the next stage, an analogue of a successful guardian spirit quest, is expeditiously reached.
7
This scene has several instances of the elementary three-step relation as well.
The three spoken statements in (A) can be taken as three steps (onset, ongoing, outcome)
of the initial plan. First the boy is addressed (with the transitive verb-stem -lxam); then
the people are spoken to generally, broadcast (with the intransitive stem -kim): then some,
not all, are addressed, as indicated by -lxam instead of -kim, presumably excluding the
boy. After the first stanza, which has everyone in place, three stanzas each have changes
of location with the onset, ongoing, outcome pattern. The people cross, go on, arrive at
the reeds (8, 9, 10); they run off, go straight home, none are left (17, 18-19, 20-21); the
boy starts, follows, arrives running (28, 29, 30). Such a three-step change of location
constitutes all of (C) and (E). (B) and (D) overall have three-step sequences, but not of
movement as such. Reach the reeds, cut them, instruct them (B), boy calls to go home,
reeds answer, boy searches in vain (D).
The entire translation is given as an appendix below, because it will figure in
other parts of this essay as well. This version replaces that in chapter 4 of Hymes 1981.
8
Benedict herself remarks of the tales she recorded: They give the literary style
to which all the stories in Cochiti conform but which can never be completely reproduced
without recording the text (xiii). Her relative confidence about style probably was based
on the fact that Franz Boas recorded a number of tales in the language itself, and
published a characterization of it (Boas 1928). The translations of the stories he recorded
in text are included in her monograph.
been re-elicited (Kari 1981:1-15). One is the widely known story of Raven
obtaining the light of the sun (Chapman 1914:22-26, 109-15).
The patterning of verses and scenes uses relations of two and four,
and the story as a whole has four acts. The four acts integrate two distinct
plots. Each involves three acts (Hymes 1990-94).
In the first plot, Act I establishes that no young man can marry a
certain woman, Act II has Raven succeed in entering her, Act III has her
discovered to be pregnant and Raven born as her child. The woman who
would not marry has been overcome.
In the second plot, Act II introduces Raven, who flies in darkness, Act
III has him born to the daughter of the man who controls the light, Act IV
has him make off with the light. The world has been set right.
In the narrative as a whole, Act I involves the young woman, but not
Raven. Act IV involves Raven, but not the young woman. Each is in three
acts, and they share the central acts II and III.
In the Cochiti narrative, there are two scenes. The first is about
Coyotes attempt to imitate a bird, the second about what happens after he
fails. Both scenes involve interplay of relations of three and five with
relations of four, but differ in internal form.
The first scene has three stanzas, the second four. In the first scene
the first and third stanzas each have four verses. The first elaborates pairing
in each verse in terms of opposition between what is high (a bank of paper
bread) and what is low (a pond of sweet-corn milk). The first pair of verses
have to do with what is there, the second with what Crow does (sing, then
bite and fly down to drink). The third stanza also has two pairs. Coyote eats
and wishes to drink in one, prepares to jump and jumps (to his death) in the
other. The middle stanza has five verses. They interlock with Crows song
as pivot. Coyote comes along and asks for the song, Crow agrees, Crow
sings. The outcome of one three-step sequence is onset to a second: Crow
sings, Coyote listens and learns, says he is ready to start.
Only after long consideration of this first scene did I realize that it is
analogous to the second, if the two interlocking sequences are counted
together with the stanzas on either side. In the first stanza there is only
Crow, in the last stanza only Coyote. In the two interlocking sequences
there are both Coyote and Crow. Three for Crow, then, and three for
Coyote, in a series of four.
This interplay of three and four complements an obvious interplay in
the four stanzas of the second scene. Crow takes Coyotes eyes herself, then
340 DELL HYMES
summons those who use fur, then those who eat meat. Finally an old man
comes and takes the bones for soup for his wife. The first three stanzas
show Crow in charge, the last three are about the use of Coyotes body. The
first stands apart from the practical uses of the rest, because Crow simply
plays with the eyes, shaking them so that they sound like bells. (An
audience would recognize a popular incident, often the frame of an entire
story, in which a bird takes Coyotes eyes.) The last stands apart because
the old man comes without reference to Crow. The second and third stanzas
belong to both sequences, involving both Crow and usefulness.
The storys two scenes are alike in beginning with Crow and ending
with Coyote, each having a three-step sequence that overlaps the sequence
of the other. They differ in expressive shape in ways that further analysis of
Cochiti may clarify. It may be accidental that these two instances of
interlocking sequences of three in a set of fourone from Alaska, one from
New Mexicoare the only ones known. The device may not be as rare as it
now seems.
Ethnopoetics and Editing
If organization in groups of lines is pervasive in oral narrative, then

the editing of oral narratives for publication should take that into account.9
Indeed, any presentation of a narrative on the page implies a hypothesis as
to its form (cf. Hymes 1987). Yet it is still possible to encounter oral
narratives presented as block paragraphs (see Shes a widow above).
Often, to be sure, oral narratives are presented as sequences of lines,
carefully transcribed and edited (Tedlock 1972, Wiget 1987, Parks 1991),
but the possibility of organized relations among the lines is not considered.
Even when verses are identified, relations among them may not be
(Kroskrity 1985).10 In general, we should realize that complex artistry in
9
The communities from which come the South Slavic epic poetry studied by Lord
and Foley very likely also tell unsung stories that make use of ethnopoetic patterning; see
Foley 1995. L.D. Perkowski (1993) has shown its presence in a series of recently
collected Bulgarian narratives.
10
For the organization of lines in a Zuni text published by Tedlock, see Hymes
1980, 1982. Tedlocks response (1983:56-61) seems not to allow for the possibility of
relations not auditorily perceptible. For rhetorical relations among verses, stanzas, and
the organization of lines may be natural to users of language, and flourish

wherever language does.
Presentation in terms of lines and verses makes visible the shaping
artistry of narrators, all that complex wealth of minutiae which in their line-
for-line aspect we call style and in their broader outlines we call form
(Burke 1968:38). The reading is slowed, which makes it far more possible
to perceive repetition, parallelism, and succession in the particular text, and
what is constant and variable among texts (cf. Hymes 1981:ch. 6; 1985).
Such analysis contributes to a general theory of the competence and
practices involved in oral narrative itself.
Verse Analysis and Manuscripts Interact
Relations among verses interact with the details of manuscript

sources. Manuscript evidence may clarify what is otherwise puzzling about
such relations. The presence of such relations may indicate the integrity of
an original source, and the failings of a published one. It is fair to say that
all the published sources for Native American narratives need to be
examined, and re-edited, in the light of the original manuscripts and verse
analysis for the choices and changes that have been made.
The Deserted Boy. This text, dictated to Edward Sapir in 1905, has
three instances. The third leads to a discovery in the one text that is
remarkable for poetics comparatively. (As indicated above, the story is
given in full in an appendix, because of its importance to more than one part
of the paper).
(1) The first lines of the story were published as follows (using the
published translation at this point):
Some time long ago the (people) said to the boy:

The boy was considered bad.
So then they said....
scenes, disclosed by quotative particles in a Hopi performance, see Hymes 1994c. I have
sketched the verses and stanzas of the first text in Parks 1991, Alfred Morsettes How
Summer Came to the North Country, and have prepared an account of the stanzas and
scenes in Dewey Healings Bird Story (Arizona Tewa) presented by Kroskrity 1985
(cf. Kroskrity 1993).
342 DELL HYMES
The field notebook shows that Some time long ago does not actually start
the story. The story starts with Now then (aGa kwapt). The word
rendered Some time long ago (GanGadix) precedes The boy was bad
(mean). Nor is that fact accidental.11
The second statement in a Chinookan myth often enough describes the
character of one of the actors. Usually this is done through a characteristic
activity, understood to be virtuous or not. An actor characterized as virtuous
will not come to harm at the end. Here the boy is characterized as bad, but
the badness is displaced: long ago.
Another of Simpsons texts, one about Clothing in the section on
customs (Sapir 1909:182), does begin with this time expression, translated
there In olden times. A second paragraph in the same text (but about
tools) begins the same way. Perhaps this is why Sapir thought the
expression should be first in this story, and of course it seems right there,
given our familiar Once upon a time. In The Deserted Boy, however,
long ago has structural work to do. The boy will not end badly, but as a
wealthy hero, taking revenge. His meanness is a once, but not a future,
thing. Louis Simpson keeps faith with the convention of a statement of
character in second place, letting a hearer know that what follows upon it in
this case is the immediate action, not the final outcome.
As always, one has to take seriously the exact detail of what was said.
Formal analysis need not displace the manuscript, but may underscore its
integrity. The two together provide as sure as possible a basis for
interpretation.
(2) A second instance also has to do with a formal anomaly. The
narrator, Louis Simpson, marks verses regularly with an initial pair of
particles, translatable as Now then (aGa kwapt), as we have seen. The
common alternatives are regular too: a second sequence may have another
pair, Now again (aGa wta) instead of Now then.12 A turn at talk is
11
This example was intended to form section 5 of the original article (Hymes
1976), but was omitted from both it and Hymes 1981. Cf. 1981:163, line 15. For the
symbol G, note 12 below.
12
In Wishram words C is used for voiceless affricate (English ch), E is used for
schwa (like the vowel in English but), G is used for a voiced velar stop, L for a voiceless
lateral fricative, S for a voiceless shibilant (English sh), x for a voiceless velar fricative
such as in German Ich, X for a voiceless velar fricative such as in German ach. A
always a verse, however begun. Simpson builds stanzas and scenes again
and again with sets of such verses. At one point in The Deserted Boy,
however, this regularity fails. Nor can it be taken to have expressive point.
Where there ought to be a third pair, there is just one particle, then. Some
narrators do use this single particle as a marker of verses, but not Louis
Simpson.
Sapirs field notebook III, pp. 94-97, shows that at this point of formal
irregularity there is an irregularity in transcription. The words of one line
are inserted above the words of a line that follows. Either the inserted words
were initially missed by Sapir, who went back to write them, or they were
retroactively supplied by Simpson. The latter seems more likely. The verse
with a single particle completes an expected sequence of three; the
discrepancy suggests recovery in the midst of distraction. (Hymes
[1991:156-58] indicates what the content and context suggest was intended.)
(3) The third example involves recognition of conventions of
patterning that had been missed. In the final act, the published text has the
following five lines (published, of course, as prose):
Now snow, lightly, lightly.
There is no food among the people,
the people are hungry.
Now then the people said,
Let us go to the boy.
That is a reasonable sequence. The field notebook, however, shows that for
publication Sapir changed the order of the lines. If the order in the notebook
is identified as abcde, then the printed lines are in the order ecdab. The
change appears to be an interpretation. The field notebook shows no
insertions. What it does show are carets and parentheses. These indicate
transposition in two steps. This fact, and the fact that the translation remains
continuous in the original order, suggest a result of editorial attention, not of
interaction with a narrator.13
When the relevant lines are considered in the order in which they were
written down, and presumably spoken, they lead to reconsideration of the
organization of the act as a whole. One gains a richer sense of the ways in
which initial particles are used, of their motivation and consistency, a
consonant followed by is glottalized.

13
See Hymes 1981:161 for details.
344 DELL HYMES
further confirmation of traditional referentiality and the premise that

poetic meaning depends fundamentally on poetic structure (Foley 1991:6,
14).
Here is the notebook order:

Let us go to the boy.
Now snow, lightly, lightly.
Notice that the published order puts the last line at the beginning. This may
be because it has initial now. A single initial now (aGa) is sometimes
used by Louis Simpson. Indeed it is used in each act of this narrative, but the
circumstances are different and revealing.
(a) The last line (30) of the first act is Now, no people. That sums
up the outcome of the desertion of the act, and the condition of the act to
follow. (This now, however, does not mark a new verse. It does not begin
a predication, but completes one. See note 18 below).
(b) The first scene of the second act ends with lines each beginning
with a single now, two of them. These lines conclude the fifth of a strict
sequence of verses. The boy fishes five times. Four times we are told that
he has caught one (two, three, four) fish, eaten half of what he has caught,
and saved half for the morning. The first time begins with Now then, the
four that follow with Now again. The fifth time we are not told what he
has caught; rather:
Now five times the boy had fished.

Now he had become a grown man.
A sequence of five is a standard pattern that arouses expectation of

completion, but the expected completionwhat he has caughtis held over
for a scene of its own, an extravaganza in which the boy, discovering that a
being in the river has given him, not fish, but prepared winter food, sings
and waves a feathered cloak. Three of the verses indeed begin with the
emphatic particle quStaxa behold! indeed!
The lines at the end of the fishing scene sum up what has occurred (he
has fished five times), and what will be the condition of what follows (he has
become a grown man).
(c) The third scene of the second act ends with five lines (116-20) that
each begin with a single now. Lacking a following then, the onward
push of the narrative is suspended. There is a moment of lyric unity between
the boy and the woman who comes to him. (Such moments for a man and
woman occur, variously marked, in a Clackamas narrative from Virginia
Howard and one Kathlamet from Charles Cultee). The lines culminate and
sum up the reward of what the boy has done (food, a wife, power). The food
and power are a condition of what is to follow. (In Victoria Howards
Clackamas version, so is the wife).
(d) The last line (167) of the story is Now only the two old women
remained. It sums up the outcome of revenge.
(e) In the notebook order of the five lines in question, Now snow,
lightly, lightly occurs at the end of a stanza (III (B)). It does not sum up a
state of affairs, but it anticipates what is to follow. Perhaps in this respect it
complements the other instance in Act III. The uses of a single Now at the
end of a unit in Act I (31) and Act II (78, 79; 116-20) both sum up and
anticipate. In Act II one anticipates (31), the other sums up (167).
Another pattern intersects this one. There are three mentions of
snow in the narrative. In each of the others snow is the third element in
a sequence of three lines.
Then now he raised the east wind,

the east wind became strong,
and it snowed
Now again he treated them this way,
a strong east wind blew,
moreover now there was snow.14
It seems reasonable to take the first mention of snow as participating in

that pattern. The people who abandoned the boy twice drown in the midst of
wind and snow. Here the condition of that outcome, snow, has begun.
(4) Couplets: Act III. Notice that the two lines preceding the first
mention of snow (141-42) are odd in terms of Chinookan patterns of verse
marking. What precedes is marked as a verse by initial Now then and a
turn at talk (139-40). What follows is marked as a verse by initial Now.
14
Lines 152-55, and 162-64. Line 153 is an English explanation that is not part of
the narrative proper.
346 DELL HYMES
What intervenes has no initial particle, no turn at talk, yet it seems to have
the position of a verse.
One might think of the lines There is no food among the people, / the
people are hungry as part of a preceding verse, Now then the people said, /
Let us go to the boy. But Chinookan quoted speech always is the end of
the verse of which it is a part. If the two lines in question are to be a verse,
they ought to begin with a marker. Instead they begin, literally No-thing
food people-at.
It has taken me twenty years to notice that the two lines are a couplet,
a semantic couplet. Each says much the same thing:

With this recognition it was a matter of a moment to consider two other lines
about people as a couplet as well:
All died in the water,

the people were drowned. (156-57).
So to consider these lines was to provide an answer to longstanding

dissatisfaction with the form of the act. The way I had published it, after
much wrestling, had never seemed quite right, and I had tinkered with it in
the interval. Lines that should be structurally parallel were not. Now they
could be. Indeed, now the recognition of lines 156-57 as a couplet, and
thereby a single unit, seems inescapable.
The context is this. Line 144 is strongly marked as an onset,
beginning as it does with three particles in a row, Now then again.... A
few lines later Then now is strongly marked as an onset, as an inversion
of the usual sequence, Now then. If each is the beginning of a stanza,
then each also ends in parallel fashion. At the end of each the boy recalls of
the people, they abandoned me. Such coming round to the same point is
an important device in the tradition. These two sets of lines, then, make
perfect sense as stanzas, with strongly marked beginnings, parallel endings
essential to the theme, series of verses, five and three, fulfilling a pattern
numberif and only if All died in the water, / the people were drowned,
the people couplet in the second of the two (156-57), is a structural

element.
As with the preceding couplet, these lines express a common theme in
varied form, and stand apart from what precedes and follows. What follows
counts as a separate verse because it is a turn at talk (and thematically
parallel with the ending of the preceding stanza). What precedes is itself
formed on a model repeated in the stanza that follows: he did this, a strong
east wind, snow (162-64). (I set the lines apart in the earlier analysis, but
did not reach the point of counting them as a verse).
Act III, then, has three instances of a three-step sequence ending with
snow, and three instances of a psalm-like pair of lines involving people
(141-42, 156-57, 165-66). Recognition of these patterns makes possible a
coherent, pointed shape for the act as a whole.15
As said, I missed this shape in my published article on the story. (To
be sure, it was one of the first texts I analyzed in terms of verse patterning).
To discover the original order of the lines involving Now snow, as I had
done, was not enough. I was not intimate enough with Louis Simpsons
style, not sure enough of its constants and those of other Chinookan
narrators. Not recognizing the structural role of these two patterns (a triad
ending with snow, couplets), I could not reconcile the different kinds of
repetition and marking in the act with an overall expectation of three- and
five-step sequences.16
Now a clear working out of implicit narrative logic, explicitly
marked, can be seen. The first five verses form a stanza with interlocking.
The third verse, the grandmothers crossing to the boy, is outcome of the
preceding two, and onset of the two that follow. The next verses can be
seen as a sequence of three pairs of verses. Such sequences are common
enough in Chinookan narrative. The implicit rhythm of expectation within
each pair is this, then that. The first pair of verses (131-32, 133) have to
do with the two old women: they go across, they are there a long time. The
second pair (134-38, 139-40) have to do with the news and what is said:
there is much food at the boys, let us go across. The third pair juxtapose
15
For all the features of the act, please see the appendix, which replaces the text
presented in Hymes 1976 and 1981:ch. 4.
16
See Hymes 1981:159-64, for the earlier consideration. These pages and others
cited above are captioned Structural philology (a) and Structural philology (b).
348 DELL HYMES
the couplet: the people are hungry, and now there is snow. A rhythm of
this, then that joins irresistible motivation (no food) to incipient danger.17
The third stanza is the peripety: the grandmothers come, they get close
to the boys house, other people start across. The boy turns, looks, sees.
Doing so, he echoes the triplet in which he discovered the fire his
grandmothers left him, and remembers his abandonment. By implication, he
resolves how to act.
This memory is doubled (stanzas C, D), and so is the drowning of the
people (D, E). All this is part of an interlocking relation among the five
stanzas. The first two stanzas (A, B) have the presence of food at the boys
discovered. The last two stanzas (D, E) have the people who come for it
destroyed. The middle stanza (C) has the people start across and the boy
resolve. That is the outcome of one three-step sequence (discovery, wider
discovery, confrontation) and the onset of another (confrontation, outcome,
further outcome).
The texture of the scene includes other three-part relations as well.
The grandmothers cross three times (A, B, C). Snow comes three times (B,
D, E). There are three couplets about the people (B, D, E). Each of the last
three stanzas (C, D, E) actually ends with the theme of the abandonment,
two with memory of those who did abandon, the third with the safety of the
grandmothers who did not.
(5) Couplets: Act I, II. Such couplets occur in each act. In Act I they
have to do with the peoples abandonment of the boy:
straight across they ran,

straight across they went (18-19)
not a person on this side,

all on that side (20-21).18
17
It is possible to take the stanza as five interlocking verses, since the first three
verses make sense as a three-step progression of onset, ongoing, outcome (with
traditional reference to other versions in which how the news gets out is spelled out).
The third step, becoming news, might in turn be the onset of another three-step
progression (there is food at the boys, let us go, now snow). But that would ignore the
lines of the couplet, which have no normal place in any of the five verses.
18
Lines 30-31 he arrived running: / Now, no people might seem a couplet from
the standpoint of counting lines. To take it as a unit would give the stanza three elements.
What we have here, however, is the conjunction of two other narrative patterns: the first
The parallelism of the lines was readily seen and expressed from the start,
and the organization of the act is not affected by counting the pairs as single
units.
In Act II, on the other hand, the recognition of couplets forces
recognition of relations that had been ignored. The fifth stanza of the first
scene is clearly strictly patterned in terms of going to fish five times, so
much so that presenting it as just that seemed obvious. But if the last two
lines are a couplet, and hence a unit with the status of a verse, matters are
different. If lines 77 and 78-79 are a pair of verses, what precedes them does
not fit in a consistent pattern with them, unless also consisting of pairs. And
of course it does.
In any other narrative sequence of successive days, the occurrence of
morning, let alone again morning, that is, of initial markers for
recurrence and a new point in time, would have automatically been seen as
marking a new verse. Here the obvious sequence of five days induced a
false security, and the lines about eating the next day were tucked in with the
catching. Five days, five verses.
Now it is evident that the stanza is expressively elaborated with not
five verses, but five pairs of verses. The first four pairs have fishing one
morning and eating the remaining half the next. The fifth pair has going the
fifth time, and a dramatic change of perspective in a concluding couplet, the
sudden disclosure that all along the boy had been achieving adulthood (78-
79).
I know no other instance of such narrative couplets in the region. Such
may be found, but at present it is impossible to think in terms of diffusion.
Perhaps the couplets are an indigenous development of the pairing that is
widespread in the three- and five part-patterns of the region, often to
highlight a focus of action. They can be seen as an intensification of it. I
have no hypothesis as to why they occur only here in what is known of
Louis Simpsons narratives. They may be a sign of how much it meant to
him to etch with decisive strokes, as a triumphant guardian spirit quest, the
story of an abandoned boy.
three lines are an example of the common three-step pattern of action (onset, ongoing,
outcome): he started home, he followed behind, he arrived running. The third and fourth
lines are an example of an action coupled with an object of perception: he arrived
running; now, no people.
350 DELL HYMES
Significance of what is missing: Salmons Myth
Manuscripts may show a published text to be missing a line or two,

and restoration of the missing line(s) may show the structure of the narrative
to be different (cf. Hymes 1985:406-7, Hymes 1994b). In one case missing
lines reinforce interpretation by the fact of being missing where they are.
In the last decade of the last century Franz Boas searched for speakers
of the Chinookan languages spoken near the mouth of the Columbia River.
He found Charles Cultee, with whom he intended to stay a day or two, but
whose intelligence and ability caused him to return several times. Cultee
was the only person from whom Boas could obtain connected texts in either
Kathlamet or Chinook proper (which I call Shoawalter to distinguish it).
Wanting to check the accuracy of Cultees command of Kathlamet, Boas
asked him in 1894 to tell again two stories he had told in 1891. With one,
War of the Ghosts, he got a variant about people on the other side of the
river. With the Salmons myth, he got a version elaborated in the service
of a theme.
Both versions have two parts. In one Salmon returns up river in
spring, and is hailed five times by plants along the bank. They insult him
and assert that (in his winter absence) the people would have starved if not
for them:
At last my brothers son arrives,

the one with maggots in his buttocks.
If I were not a person,
your people would have died.
Salmon shows no offense, but recognizes each plant as an aunt or uncle,

gives it a gift, and places it where it will be in times to come.
In the second part Salmon and his company meet three people coming
down the river toward them. They claim to have gone all the way upriver to
the Cascades and be returning in a day. The leading person is a woman.
Her spokesman implies the truth of the claim by speaking the upriver
language, Wasco, and naming in Wasco (untranslated) a major womans
food, camas. Salmon takes umbrage at these, twists their necks, and denies
their claim. It will take five days to reach the Cascades.
In this 1891 telling the first part begins somewhat leisurely; several
lines explain the situation. The second part begins dramatically, with
Salmon issuing directions three times in succession, and using ironic
questions. It is accomplished in five stanzas, one scene. In the 1894 telling
it is the second part that begins somewhat leisurely, as Salmons company
move on upriver. The verses are ordinary threes and fives; no dramatic pairs
of Salmons behests and responses to them, no ironic questions but questions
in the passive at first (they were asked). There are three scenes, not one.19
A second section deals with the three who have come downriver: Salmon
pronounces what they will be. And the order in which they are dealt with is
reversed, so that the last one is Flounder, whom Salmon tells to remain in
the river in the winter.
Salmon is a contested figure in terms of gender. In other narratives he
is shown as proud and peremptory with women. Victoria Howard
transforms and ultimately excludes Salmon from a version of this very story
(Hymes 1986). Here he is made to acknowledge the importance of womens
food (plants) to the survival of the people. One can see his behavior in the
second part as a result of suppressed anger at the insults he must suffer
silently in the first part. In the 1891 telling the anger is overt. In the 1894
telling it is not. Evidently the reason is the further ending. By having
Flounder be year round in the river, Salmon forever undercuts the claim of
the plants to be the only winter source of food.
The field notebook makes a minor difference to the number of lines in
the 1894 telling (one notebook line appears to have been missed in the
printed text). What is telling for interpretation is the fact that each time
Cultee skipped a line in the scene just before the second part. In the first
telling he went right on. In the second telling, so the notebook shows, he
remembered the omission and inserted it a moment later. What Cultee did is
invisible in the printed text, because in editing Boas put the remembered line
where it should have been.
19
The relations given in Hymes 1985 should be revised as follows:
[i] [Encounter] (A)92, 93-94, 95
[ii] [Colloquy] (A)(abc) 95-99, 100-1, 102
(B)(abc) 103, 104, 105-9
(C)(abc) (110-11, 112-17, 118-23)
[iii] [Outcomes] (A)[Twisted] (abc) 124-26, 127-29, 130-31
(B)[Pronouncement] (a) 132-34
(C)[Thrown] (abc) 135, 136-38, 139-43
352 DELL HYMES
The notebook indicates that Cultee was quick to get to the second part
in 1891, but not in a hurry in 1894. In both tellings, one can infer, he wanted
the second part to offset the humiliation of Salmon in the first. In 1891 he
hurried to the part in which Salmon can be in command, and dramatized that
commanding role (a marked pattern of verses, ironic questions), letting go a
line along the way. In 1894 he did not hurry, but paused to restore an
omitted line; nor did he mark the new part expressively. He had Flounder up
his sleeve.
The two tellings convey a common concern on Cultees part. The
notebooks underscore it. Differences in response to a slip in performance
covary with different ways of accomplishing a purpose.
Editing and Value
Discovering Cultees handling of omissions, discovering Louis

Simpsons ordering of lines, are examples of recovering intention (cf.
Gorman 1989:194, discussing Parker 1984). One is concerned with what the
narrator actually said, with authenticity. That has been a primary value for
many.
In these cases the recovered intention supports a form of the text that
has greater aesthetic value, if, as I believe is the case, there is aesthetic value
in the shape the narrators have given what they say. But what gives value is
not always obvious or agreed upon. Folklorists sometimes conflate versions,
choosing what appears a better passage or wording from each (eclectic
editing). If each version has its own shape, however, the result may be a
mixture partly without shape. Suppression of a line may suppress indication
of a verse; addition may add one. Either may distort a local configuration
and produce puzzling irregularity. From the standpoint of verse analysis,
such a practice is to be shunned.
To be sure, a particular performance may be both authentic and
inferior. Here is where a value other than aesthetic enters. Verse analysis is
analysis of language, and contributes to linguistics as well as to folklore,
anthropology, and literature. Noam Chomsky has led many linguists to
consider it their concern, at least in principle, to analyze, through language,
the abilities that underlie it, competence. For abilities in a broad sense,
beyond grammar, the term communicative competence has been adopted by
many (cf. Hymes 1974, 1984). Imperfect narrations may shed light on the
competence that underlies narrative, on how it works.
Even with splendid narrations, aesthetic value and analysis can easily
be at odds. Being unfamiliar with the conventions of another tradition, or
unconscious of effects deployed in our own language, we may need to have
what goes on called to our attention, pointed out, in order to see it. Where
alternative interpretations of form are possible, the alternatives must be
shown in order to be discussed. If analysis is to contribute to understanding
of competence, it must be explicit. For all these reasons, narratives must be
presented in a format that makes their analysis recoverable and clear.
I call this showing the bones of the narrative. There is analogy to
an edition of Gilgamesh that presents precisely what is there on a certain set
of tablets, as distinct from a translation that presents a continuously readable
story (cf. Kovacs 1989, Sandars 1972). In some cases, it is clear that one is
displaying relationships that, though marked, are not salient in the flow of
words, what might be called the flesh (see Hymes 1994c).
At this stage of our knowledge of many traditions, such as those of
Native Americans, showing the bones is required. When what is there is
not yet publicly known, it must be presented first. After that, surrogates of
all kinds, retellings, imitations, dramatizations can proceed. But bones come
first. To do otherwise would be to regard Popes Iliad as Homer, Lambs
Tales as Shakespeare, and Bible stories for children as Genesis, Job, and the
Gospel according to John.20
Recovery of the Old
This concern is linked to the notion of repatriation. The notion has

come to the fore in connection with the recovery of burials and other
objects taken from Native American communities. There is a textual
parallel. For many Native American communities, texts in the traditional
languages are no longer told. What remains is what has been written down.
Important as it is that Native Americans speak for themselves, texts do not.
The relations of form and meaning explored by verse analysis are like other
20
These considerations are an instance of the general issue raised by McGann
(1983), that of the need to locate editing and literary production in their particular social
nexus. Cf. Gorman 1989:194ff.
354 DELL HYMES
relations of form and meaning in language. Mostly we are not aware of

them. Analysis is necessary to make them explicit. It is a kind of
repatriation, then, for those of us fortunate enough to be able to do so to help
recover in older texts their lineaments of shaping artistry.
This may go against the grain of a focus on performance and theory of
a certain kind. When I spoke about such work some years ago at the
Smithsonian, using a text from a now extinct language of Oregon, someone
asked why work with such (limited) materials, why not work with materials
in which one can hear and see the performance? The short answer is that I
am from Oregon. It matters to me, and to people I know, to recognize the
value of what textual record there is. From this standpoint, recovery of the
old as such matters. A few scholars are pursuing this kind of work. Let me
illustrate its value with a few examples.
Multilingual source? About a century ago Franz Boas recorded some
stories from the now extinct Salish people known as Pentlatch. Some exist
now in manuscript in Pentlatch, some in published translations in German,
but not all in both. It is likely that some narrations in Chinook Jargon were
translated directly into German. In any case, Kinkade (1992) is able to
clarify the relation between the two kinds of source, comparing a manuscript
text in Pentlatch and its published German translation with the help of verse
analysis.
Recovering verbal play. Berman (1992) provides a notable example
of recovering the value of a text. She notes that it is not the original texts in
Kwakwala (Kwakiutl), but Boas translations of them, that have become
the primary source for generations of scholars. Berman observes (157):
Lvi-Strauss to the contrary, the meaning of a myth lies within the

narrators use of language, not outside it. Boas knew this, which is why
he left us eleven volumes of Kwakwala texts. If Boas translations to
those texts are unreliable, I believe it is at least in part because he did not
intend for them to be relied on. For Boas, the texts were in and of
themselves the end products of ethnography, and the translations a
necessary evil, an aid to those without fluency in Kwakwala.
Berman herself commands the language, and sources scattered over a

number of years, so that she is able to reconstruct choices that Boas made,
not only in translation, but also in composing a dictionary. She is able to
show that a text is couched in verbal play that escaped Boas and that has
escaped everyone since.21
There is in general a need for anthropologists and folklorists to
understand their field as philologyto return to manuscript sources, to
discover what has been excluded, rearranged, normalized, misunderstood
(cf. Foley 1992:276, 290). What can be known can be expanded in the
archive as well as in the field.
Performance register (1). The manuscript sources of Boas two
volumes of Chinookan texts (Boas 1894, 1901) show an allegro style of
dictation. Boas appears to have normalized elisions and published full
forms. The easy style suggests that the narrator, Cultee, was not much
affected by the process of dictation, and that something of a relatively
spoken style can be recovered. That is the good news. The bad news is that
the published texts can not be confidently relied upon until the uncorrected
originals are studied. The sources of some titles and incidents, published in
the language by Boas, have not yet been located in the notebooks. (There
are also many supplementary verbal forms, never published, which I did not
learn about until I had written a dissertation grammar on the basis of the
published material alone).
Performance register (2). Even with narratives told in English, the
English style of the narrator has probably been revised. Here is one scene
from a narrative in Tillamook Salish which has attracted attention.22 There
are four stanzas, separated by space. Verses begin flush left. Closing
21
Berman does not actually indicate the verses in the text, only the two parts to
each stanza. Verses can be recognized in terms of the initial element llai then (pp.
130, 131-32) and turns at talk. Stanza (A) has two verbs of saying in its first part, Then
twice initially in its second part. Stanza (B) appears to be marked by having four framing
verbs of speaking, the first of each pair with initial Then, but then a third initial llai
and a fifth framing verb (of singing). These lines (14-17) are the peripety and the only
song. Stanzas (C) and (D) resume even-numbered patterning. (C) has initial Then and
a turn at talk with a verb of saying, while (D) has twice initial Then.
Carrying through the verse analysis, and showing it in translation, (as Berman
does in other work) brings out the special status of stanza (C). The peripety is marked in
form against the background of the rest.
22
E. Jacobs 1990, with an appendix for this story by myself; the analysis into
verses is slightly revised here. Cf. Hymes 1993b and Seaburg 1992. I am indebted to
Seaburg for the notebook original.
356 DELL HYMES
braces indicate pairs of verses that go together in a pattern of three such

pairs.
Later on his sister said to him,

You are getting grown now,
you should hunt a woman for yourself.
You are old enough to get married.
Any old thing, a dead person, is perhaps better than no wife at all.
Huh! I can do that all right, sister.
He went to look for a wife.
He returned late at night.

His sister was already in bed
and did not see him.
Presently she heard him say,
Oh! My wife is sticking me with her scratcher.
His sister thought,
Why, he must have a maiden bathing after her first menstruation.
Daylight came.
The sister arose
and built the fire.
Split-His-Own-Head got up,
he had no wife.
Where is your wife?
his sister asked. }
In bed.
Is she not going to get up? }
He told her,
No. You told me to obtain a dead person for a wife.
That is a dead woman I went and got.
She said to him,
Now you take that dead body
and put it right back where you found it. }
He took it back.
Here are the words in the field notebook (in verse analysis):
Next, she told him,

Youre getting big enough now,
you can hunt yourself a woman,
you can get married.
Any old thing, a dead person.

Huh, I can do that, all right, sister.
He went to look for a wife.
He came back in the night.

His sister was already in bed,
and didnt see him come.
Presently she heard him say,
Oh! My wife is sticking me with her scratcher.
His sister thought,
Oh he must have found a maiden
just bathing after her first mensis.
Daylight came.
The sister arose
and built the fire.
He got up,
he had no wife.
Wheres your wife? }
In bed.
Isnt she going to get up? }
No, you told me to get a dead person for a wife.
Thats a dead woman I went and got.
Oh you take that dead body
and go put it back where you got it. }
Most changes are the sort a teacher would make to dress up spoken
style for appearance in print: eliminate contractions, substitute returned for
came back, obtain and found for get and got. The expansions in
the fourth and fifth lines, like substitution of proper name for pronoun in the
third stanza, are evidently to make sure the reader does not miss the point. A
third kind of change, found in another scene, eliminates direct naming of
body parts and functions. Such changes are probably widespread in what
one is invited to read as a native voice: written norms, explanations,
propriety. But unedited wording has more the flavor of a told story, and
sometimes shows a different number of lines and local shape.
Order of narration. Presumably fundamentalists and higher critics
alike recognize that the order in which Pauls letters appear in the New
Testament is not an order he gave them, or the order in which they were
written, but editorially determined by length, longest first, shortest last.
Students of Native American collections may forget that the order in which
358 DELL HYMES
myths and tales appear is not likely to be the order in which they were told,
and that inferences based on the published sequence are suspect.
Recovering actual order can indicate something about style and interaction.
The order in which Victoria Howard dictated Clackamas texts to
Melville Jacobs in 1929 and 1930 indicates two ways in which her style
changed.23 On the one hand, the earliest recorded narratives show a great
deal of pairing of verses marked initially by now (aGa). That drops out to
be replaced in favor of far less pairing and far less explicit marking of verses
by any initial element. On the other hand, it is only a certain distance into
the relationship that she begins to end a narrative with the formal close
Story, story (kni kni). The first change seems to indicate that she was
used to a style in which two- and four-part relations were very prominent, a
style not otherwise known in Chinookan, and which she may have
experienced in hearing Molale (which she knew) or some other language at
multilingual Grande Ronde reservation, where she was born and grew up.
The second change seems to reflect a growing confidence in her narratives
as complete. (Various comments show awareness of some narratives as
incomplete.) Both changes may reflect also a growing ease in her
relationship with Melville Jacobs.24
Coos Bay: Repeated tellings. Let me end with a few lines from an
obscure manuscript that are for me a sign of grace. I have been working on
a collection to make visible to others the pervasiveness there of this kind of
poetic structure in the words of Native Americans of the Northwest, and hit
upon the title, River Poets of Native Oregon. Two years ago, just as my
wife and I were setting out for the coast of Oregon, we picked up a
forwarded letter from a man we did not know. He was director of cultural
heritage for the Confederated Tribes of Coos, Siuslaw, and Lower Umpqua
Indians, he knew we had visited such people many years before (the first
summer of our marriage in fact), experience had taught him that linguists
did not answer his letters, but how about it? We went, and found that he
had patiently assembled every known bit of documentation of the
23
See M. Jacobs 1958, 1959 and the footnotes therein.
24
The order of the published Wishram Texts (Sapir 1909) is not identical with the
order of the notebooks. The Coyote cycle is interrupted by part of the Salmon Myth, and
the moving observation printed at the end of the cycle does not occur there in the notebook.
In Clackamas Chinook Texts the last section of a myth important for its performance
sequence (Hymes 1986) is taken from a separate comment on a different notebook page.
languages, cultures, and histories of these people (including an old letter of

mine). In the course of collaboration last summer he sent me a xerox of the
field notes of Harry Hull St. Clair, 2d, who in 1903 had recorded texts in
Coos that had later been published by another Boas associate who had
worked with the same man (Frachtenberg 1913).
Scrutiny of the manuscript discloses that it contains two unpublished
texts. Each is an earlier version of a text that was published. St. Clair
recorded two versions of a text entitled The Country of the Souls, and
Frachtenberg published the second (1913:no. 23). St. Clair recorded a
version of The Ascent to Heaven, but Frachtenberg obtained a fuller
version and published that (1913:no. 3). The unpublished versions have
details not present in the versions published. As in many cases, so little of
Coos tradition is known to us that details are precious. And in these cases
there is the opportunity to compare tellings (performances) by the same
narrator.25 The opportunity has remained unknown throughout most of the
century, and comes to light now through the efforts of a man of Coos
descent who has made himself a scholar.
Coos Bay: River Poets of Native Oregon. But the special serendipity
has to do with a notebook page preceding the narratives. On page 25,
numbered lines 8-12, St. Clair wrote down a few sentences that seem to have
been volunteered by Jim Buchanan, perhaps elaborating in answer to a
question. The sentences are eight in number, and group in sets of four (as
one would expect in Coos oral narrative). They seem a perfect epigraph for
a collection conceived as representing river poets of native Oregon.26
Thats the only way theyve been talking.

They didnt come from any place.
That was their only place.
They didnt know where they came from.
Every stream has people on it.

Thats how they all had a stream.
25
Sapirs field notebooks for Wishram Chinook contain an unpublished version
from the same narrator, Louis Simpson, of the first published myth. The degree to which
there is something like formulaic recurrence could be established.
26
Buchanan spoke in Coos and then provided a translation, written down by St.
Clair word by word below the Coos. The last words of line 3 and line 8 are the same in
Coos, their land, earth, country, ground, place.
360 DELL HYMES
Thats the way they know themselves.

All other tribes had their stream as their land.27
University of Virginia
References
Bauman and Briggs 1990 Richard Bauman and Charles L. Briggs. Poetics and
Performance as Critical Perspectives on Language and
Social Life. Annual Review of Anthropology, 19:59-88.
Benedict 1931 Ruth Benedict. Tales of the Cochiti Indians. Bureau of

American Ethnology, Bulletin 96. Washington, D.C.
Reprinted Albuquerque: University of New Mexico Press,
1981.
Berman 1992 Judith Berman. Oolachan-Womans robe: Fish, blankets,

masks, and meaning in Boass Kwakwala texts. In On
the Translation of Native American Literatures. Ed. by
Brian Swann. Washington, D.C.: Smithsonian Institution.
pp. 125-62.
Boas 1894 Franz Boas. Chinook Texts. Bureau of American

Ethnology, Bulletin 20. Washington, D.C.
Boas 1901 ______. Kathlamet Texts. Bureau of American Ethnology,

Bulletin 26. Washington, D. C.
Boas 1928 ______. Abstract Characteristics of Keresan Folk-tales.

International Congress of Americanists, 20, i:223-24.
Burke 1968 Kenneth Burke. Psychology and Form. The Dial, 79, i
(1925):34-46. Rpt. in his Counterstatement. New York:
27
This paper was prepared for the spring 1993 meeting of the Society for Textual
Studies. I want to thank John Foley for inviting me to take part. Since the meeting I
have revised the analysis of the text, The Deserted Boy, after recognizing the role of
couplets in it, and have added comments on the recognition of lines in Anglo-Saxon
studies. I want to thank Nick Doane and Joe Russo for their stimulating papers at the
meetings, and Hoyt Duggan for his encouragement.
Harcourt, Brace, 1931. 3rd ed., Berkeley: University of

California Press, 1968.
Chapman 1914 John W. Chapman. Tena texts and tales from Anvik,
Alaska. Ed. by Pliny Earle Goddard. Publications of the
American Ethnological Society, 6. Leiden: E. J. Brill.
Foley 1991 John Miles Foley. Immanent Art: From Structure to

Meaning in Traditional Oral Epic. Bloomington: Indiana
University Press.
Foley 1992 ______. Word-power, Performance, and Tradition.

Journal of American Folklore, 105:275-301.
Foley 1995 ______. The Singer of Tales in Performance.

Bloomington: Indiana University Press.
Frachtenberg 1913 Leo J. Frachtenberg. Coos Texts. Columbia University

Contributions to Anthropology, 1. New York and Leiden:
Columbia University Press and E. J. Brill.
Gorman 1989 David Gorman. The Worldly Text: Writing as Social

Action, Reading as Historical Reconstruction. In Literary
Theorys Future(s). Ed. by Joseph Natoli. Urbana:
University of Illinois Press. pp. 181-220.
Hymes 1974 Dell Hymes. Foundations in Sociolinguistics.

Philadelphia: University of Pennsylvania Press.
Hymes 1976 ______. Louis Simpsons The Deserted Boy. Poetics,

5, ii:119-55. Rpt. in Hymes 1981:ch. 4.
Hymes 1980 ______. Particle, Pause and Pattern in American Indian

Narrative Verse. American Indian Culture and Research
Journal, 4, iv:7-51.
Hymes 1981 ______. In vain I tried to tell you: Essays in Native

American Ethnopoetics. Philadelphia: University of
Pennsylvania Press.
Hymes 1982 ______. Narrative Form as Grammar of Experience:

Native American and a Glimpse of English. Journal of
Education, 164, ii:121-42.
362 DELL HYMES
Hymes 1984 ______. Vers la comptence de communication. Paris:

Hatier-Crdif.
Hymes 1985 ______. Language, Memory, and Selective Performance:

Cultees Salmons Myth as Twice Told to Boas. Journal
of American Folklore, 98:391-434.
Hymes 1986 ______. A Discourse Contradiction in Clackamas

Chinook. Victoria Howards Coyote made the land
good. In 21st International Conference of Salish and
Neighboring Languages. Compiled by Eugene Hunn.
Seattle: University of Washington, Department of
Anthropology. pp. 147-213.
Hymes 1987 ______. Anthologies and Narrators. In Recovering the

Word. Ed. by Brian Swann and Arnold Krupat. Berkeley:
University of California Press. pp. 41-84.
Hymes 1990-94 ______. Na-Den Ethnopoetics: A Preliminary Report. To

be published by the Alaska Native Language Center,
University of Alaska, Fairbanks.
Hymes 1991 ______. Ethnopoetics and Sociolinguistics: Three Stories

by African-American Children. In Linguistics in the
Service of Society: Essays to Honour Susan Kaldor. Ed. by
Ian G. Malcolm. Perth, Australia: Institute of Applied
Language Studies, Edith Cowan University. pp. 155-70.
Hymes 1992 ______. Use All There Is to Use. In On the Translation

of Native American Literatures. Ed. by Brian Swann.
Washington, D. C.: Smithsonian Institution. pp. 83-124.
Hymes 1993a ______. Inequality in Language: Taking for Granted.

In Language, Communication, and Social Meaning. Ed. by
James E. Alatis. Georgetown University Round Table on
Languages and Linguistics, 1992. Washington, D.C.:
Georgetown University Press. pp. 23-40.
Hymes 1993b ______. In need of a wife: Clara Pearsons Split-His-

(Own)-Head. In American Indian Linguistics and
Ethnography in Honor of Laurence C. Thompson. Ed. by
Antony Mattina and Timothy Montler. University of
Montana Occasional Papers in Linguistics, 10. Missoula:
Department of Linguistics, University of Montana. pp.
127-62.
Hymes 1994a ______. Bernstein and Poetics. To appear in a volume

ed. by Paul Atkinson and Sara Delamont. Cresskill, NJ:
Hampton Press. Forthcoming.
Hymes 1994b ______. Reading Takelma Texts: Frances Johnsons

Coyote and Frog. In a volume ed. by Roger Abrahams.
Bloomington, IN: Trickster Press. Forthcoming.
Hymes 1994c ______. Helen Sekaquaptewas Coyote and the Birds:

Rhetorical Analysis of a Hopi Coyote Story.
Anthropological Linguistics. Forthcoming.
Hymes 1994d ______. Arikara Rhetoric: Ethnopoetic Suggestions. In

a volume ed. by John D. Nichols and Arden Ogg.
Forthcoming.
E. Jacobs 1990 Elizabeth Derr Jacobs. Nehalem Tillamook Tales. Ed. by

Melville Jacobs. Northwest Reprints. Corvallis: Oregon
State University Press. Orig. publ. Eugene: University of
Oregon Press, 1959.
M. Jacobs 1958 Melville Jacobs. Clackamas Chinook Texts, Part 1.

Research Center in Anthropology, Folklore and Linguistics,
8; International Journal of American Linguistics, 24(1),
Part 2. Bloomington: Indiana University.
M. Jacobs 1959 ______. Clackamas Chinook Texts, Part 2. Research

Center in Anthropology, Folklore and Linguistics, 11;
International Journal of American Linguistics, 25(2), Part
2. Bloomington: Indiana University.
Jakobson 1960 Roman Jakobson. Closing Statement: Linguistics and

Poetics. In Style in Language. Ed. by Thomas A. Sebeok.
Cambridge: MIT Press. pp. 350-77.
Kari 1981 James Kari. Athabaskan stories from Anvik. Texts

collected by John W. Chapman. Fairbanks: Alaska Native
Language Center, University of Alaska.
Kinkade 1992 M. Dale Kinkade. Translating Pentlatch. In On the

Translation of Native American Literatures. Ed. by Brian
Swann. Washington, D.C.: Smithsonian Institution. pp.
163-75.
364 DELL HYMES
Kovacs 1989 Maureen Gallery Kovacs. The Epic of Gilgamesh.

Stanford: Stanford University Press.
Kroskrity 1985 Paul V. Kroskrity. Growing with Stories: Line, Verse,

and Genre in an Arizona Tewa Text. Journal of
Anthropological Research, 41:183-99. See now Kroskrity
1993.
Kroskrity 1993 ______. Language, History, and Identity. Ethnolinguistic

Studies of the Arizona Tewa. Tucson: University of
Arizona Press.
Lvi-Strauss 1963 Claude Lvi-Strauss. The Structural Study of Myth. In

his Structural Anthropology. New York: Basic Books. pp.
206-31.
Lvi-Strauss 1987 ______. De la Fidlit au texte. LHomme 101, 27,

i:117-40.
McGann 1983 Jerome J. McGann. A Critique of Modern Textual

Criticism. Chicago: University of Chicago Press.
Parker 1984 Hershel Parker. Flawed Texts and Verbal Icons: Literary
Authority in American Fiction. Evanston, IL: Northwestern
University Press.
Parks 1991 Douglas R. Parks. Traditional Narratives of the Arikara

Indians. 4 vols. Lincoln: University of Nebraska Press.
Perkowski 1993 Liliana D. Perkowski. Bulgarian Folktales. A Study of

Narrative Form and Content. Unpubl. ms., Department of
Anthropology, University of Virginia.
Sandars 1972 N. K. Sandars. The Epic of Gilgamesh. Harmondsworth:

Penguin Books. Rev. edition.
Sapir 1909 E. Sapir. Wishram Texts, together with Wasco Tales and
Myths collected by Jeremiah Curtin and edited by Edward
Sapir. Publications of the American Ethnological Society,
2. Leiden: E. J. Brill. Rpt. in The Collected Works of
Edward Sapir VII: Wishram Texts and Ethnography. Ed. by
William Bright. Berlin and New York: Mouton de Gruyter,
1990.
Seaburg 1992 William R. Seaburg. The Americanist Text Tradition and

the Fate of Native Texts in English. Paper presented to
the American Anthropological Association, 4 December
1992.
St. Clair 1905 H.H. St. Clair II. Ms. in the Bureau of American
Ethnology, now Department of Anthropology, Smithsonian
Institution.
Tedlock 1972 Dennis Tedlock. Finding the Center. New York: Dial
Press.
Tedlock 1983 ______. The Spoken Word and the Work of Interpretation.
Philadelphia: University of Pennsylvania Press.
Wiget 1987 Andrew Wiget. Telling the Tale: A Performance

Analysis of a Hopi Coyote Story. In Recovering the
Word. Ed. by Brian Swann and Arnold Krupat. Berkeley:
University of California Press. pp. 297-336.
Wolfson 1982 Nessa Wolfson. CHP. The Conversational Historical

Present in American English Narrative. Topics in
Sociolinguistics, 1. Dordrecht/Cinnaminson: Foris.
Wolfson 1989 ______. The Conversational Historical Present. In

Analyse grammaticale de corpus oraux. LINX, 20, i:135-
50. (Universit de Paris-X-Nanterre, Centre de Recherches
Linguistiques).
366 DELL HYMES
APPENDIX
THE DESERTED BOY*
[I] [The People and the Boy]

Now then they told a boy, (A)
Long ago the boy was mean.
Now then they said,
Now you will take him for reeds. 5
Now then they told them,
You shall abandon him there.
Now then the people all went across the river, (B)
they went on,
they came to the reeds. 10
Now then they cut them off.
Now then they said,
If the boy should say,
Are you there?,
you shall answer, 15
Uu.
Now then they ran off, (C)

straight home they ran,
straight across they went,
not a person on this side, 20
all on that side.
Now then the boy, too, said, (D)

Now lets go home.
Uu,
*
Wishram words:
16 A repeated vowel symbol shows prolongation.
34 The sound of the fire is phonetically a glottalized voiceless lateral affricate; that is, t plus
voiceless 1 plus glottal stop.
81 The name of a delicacy, a mixture of dried salmon and mashed huckleberries, has a-
(feminine gender), ts, glottal stop, schwa (as in English but), and p.
96 A repeated vowel symbol shows prolongation.
110 The woman is the daughter of the spirit power who lives beneath a whirlpool. His name
has i- (masculine gender), ch, glottal stop, schwa, palatal voiceless fricative (as in
German Ich), and i, a, n.
went the reeds. 25

In vain he searched about:
no person.
Now then he too started home, (E)

he too followed behind them;
he arrived running: 30
now, no people.
[II] [The Boy, Deserted]

[i] [He survives]
Now then the boy wept. (A)

Now then he heard,
TL TL TL .
Now then he turned his eyes, 35
he looked,
he dried his tears.
Now then he saw a very little bit of flame in a shell. (B)

Now then he took that very same flame.
Now then he built up a fire. 40
Now again he saw fiber, (C)

again a little bit of it,
straightway he took it.
Now again he went to the cache,
he saw five wild potatoes. 45
Now then he thought:
My poor fathers mother saved me potatoes,
and fire was saved for me by my fathers mother,
and my mothers mother saved me fiber.
Now then the boy made a small fish-line, (D) 50

and he made snares with string;
he set a trap for magpies.
Now then he caught them.
Then he made a small cloak with magpies skin.
He just put it nicely around himself. 55
Again he lay down to sleep,
Again he just wrapped himself nicely in it.
Now then he fishes with hook and line; (E) (ab)

368 DELL HYMES
he caught one sucker,

half he ate, 60
half he saves.
Again, morning, he ate half. }
Now again he fishes, (cd)

he caught two,
one he ate, 65
one he saved.
Again, morning, he ate one. }
Now again in the morning he fishes, (ef)

he caught three suckers,
he ate one and a half. 70
Again, morning, he ate one and a half.
Now again he went to fish, (gh)

he caught four suckers,
two he ate,
two he saved. 75
Morning, now he ate all two. }
Now again he goes to fish for the fifth time. (ij)

Now the boy had fished five times.
Now he had become a grown man. }
[ii] [He sings]

Now then he examined his fish-line. (A) 80
Indeed, atsEptsEp fills to the brim a cooking-trough.
He stood it up on the ground.
Now then the boy sang.
Now then all the people watched him.
Now then they said: 85
What has he become?
Indeed! he became glad, (B)

he had caught atsEptsEp.
Thus he sang:
Atsee, atsee, 90
Ah, it waves freely over me,
Ah, my feathered cloak.
Atsee, atsee,
Ah, my feathered cloak. 95

Atsee, atsee,
Ah, my feathered cloak.
Indeed! ICE xians daughter had given him food.
Now then the boy had camped over four times; (C) 100
he camped over a fifth time.
Now then he awoke,
a woman was sleeping with him,
a very beautiful woman:
her hair was long, 105
and bracelets right up to her on her arms,
and her fingers were full of rings,
and he saw a house all painted inside with designs,
and he saw a mountain-sheep blanket covering him, both him and his wife.
Indeed! ICE xians very daughter had given him food, 110
and plenty of Chinook salmon,
and sturgeon,
and blueback salmon,
and eels,
plenty of everything she had brought. 115
[iii] [The two are together]

Now he married her.
Now the woman made food.
Now, morning, it became daylight.
Now the two stayed together quietly that day.
Now the two stayed together a long time. 120
[III] [The boy and the people]

Now then it became spring. (A)
Now then the people found out.
Now then his fathers mother and his mothers mother went straight to his house.
Now then he thought:
The two old women are poor. 125
My fathers mother and my mothers mother took pity on me in this way.
Now then he fed them,
he gave the two old women Chinook salmon
and he gave them sturgeon.
370 DELL HYMES
Now then the two old women started home, (B) 130
they went across.
A long time they were there. }
Now then it became news,

they said,
Oh! there is much salmon at the boys, 135
and much sturgeon,
and eels,
and blueback salmon.
Let us go to the boy. } 140

Now snow, lightly, lightly. }
Now then again first went his fathers mother, his mothers mother. (C)
Now then they were close to the house. 145
Now then a great many people went across toward the boy.
Now then the boy turned,
he looked,
he saw many people coming across in a canoe.
Now then he thought: 150
It was not good the way they abandoned me.
Then now he raised the east wind (D)

(there became a Walla Walla wind ),
the east wind became strong,
and it snowed. 155
All died in the water,
the people were drowned.
With a bad mind the boy thought:
This is the way they treated me,
they abandoned me. 160
Now again others went across. (E)

Now again he treated them this way,
a strong east wind blew,
moreover now there was snow.
Now again they died, 165
twice the people died.
Now only the two old women remained.
Thus the ways.

Etnopoetica Dell Hymes

Uploaded by

Copyright:

Available Formats

Etnopoetica Dell Hymes

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Etnopoetica Dell Hymes

Uploaded by

Copyright:

Available Formats

Oral Tradition 9/2 (1994): 330-370

Ethnopoetics, Oral-Formulaic Theory,

John Miles Foley (1992) has opened up a consideration of the

Constraints within and among lines

Oral-formulaic theory and ethnopoetics are both concerned with

When constraint is internal to the line, we do not hesitate to speak of

narrative, if some recurrent feature marks them as such. As already noted,

Burke writes of works accepted as literature is pervasively true of oral

...So he says, That you have to do in any house. So she says,

Here is how the bracketed passage appears when displayed in terms of

[iii] [bid accepted]

So he says to my wife, he says, (A) 60

So he says to her, he says, (B)

He says, Oh, he says, (D) 75

Got a call that afternoon. (E)

[iv] [certificates accepted]

So I go to see the house (A) 85

He says, Why? (B)

So he says, Really, we wont guarantee... (C)

So he says, She wont go for it... this, that.... (D)

So I get a call back about a day later, (E)

Okay, shes accepted.

The same relations open a narrative told to Edward Sapir in Wishram

The Deserted Boy

Here is the opening in terms of lines, verses, and stanzas:

Now then they told a boy, (A)

Now then they ran off, (C)

he too followed behind them;

not a person on this side, / all on that side.

Ethnopoetics and Editing

If organization in groups of lines is pervasive in oral narrative, then

the organization of lines may be natural to users of language, and flourish

Verse Analysis and Manuscripts Interact

Relations among verses interact with the details of manuscript

Some time long ago the (people) said to the boy:

consonant followed by is glottalized.

further confirmation of traditional referentiality and the premise that

Now then the people said,

Now five times the boy had fished.

A sequence of five is a standard pattern that arouses expectation of

Then now he raised the east wind,

It seems reasonable to take the first mention of snow as participating in

There is no food among the people,

All died in the water,

So to consider these lines was to provide an answer to longstanding

the people couplet in the second of the two (156-57), is a structural

straight across they ran,

not a person on this side,

Significance of what is missing: Salmons Myth

Manuscripts may show a published text to be missing a line or two,

At last my brothers son arrives,

Salmon shows no offense, but recognizes each plant as an aunt or uncle,

Editing and Value

Discovering Cultees handling of omissions, discovering Louis

Recovery of the Old

This concern is linked to the notion of repatriation. The notion has

relations of form and meaning in language. Mostly we are not aware of

Lvi-Strauss to the contrary, the meaning of a myth lies within the

Berman herself commands the language, and sources scattered over a