Introduction 1 3
in the classical ones: rhythmic structures may become particu-
larly highlighted at the end of content units, and such highlight-
ing may function as a signal to mark the boundaries of such units.
A further example, not contained in the texts shows a combina-
tion of repetition in inverted word order and glottalization. Chona
Dominguez. An Account of a Fire . . . Eemki pa c*emkfna ? veh.
[end of paragraph], j>a ffemkinave gemki literally: *. . . our
house, as we got burned. As we got burned, our house . . . '
To sum up we note that various features contribute — either
alone or in combination — to the marking of content boundaries.
Since the true relationship between the boundaries in the deep
structure and the features which we observe in the surface struc-
ture is not clear, we see no other alternative than to represent
carefully each of these various features. In the emphatic stress
pattern, the secondary stress [*] is written although in isolated
words it is entirely predictable. It was said before that boundary
features may even be contradictory: we may, for example, as in
Eagle Flower 62/63, have an emphatic stress pattern and the
paragraph may still go on for a further short sentence. Frequent-
ly in Cahuilla narrative style, the transition from one speaker
and his direct discourse to another speaker is marked by the fea-
ture of glottalization without the feature of pause which is noramlly
combined with it: Pa£aqaraw£h 17 . . . ? e*mem he*mu ? tum he~*
cemem . . . 'So it's you?' *Yes, it's us...* This problem will
be taken up in a larger frame of reference below 4. 3. 2.
After what has been said, the reader will regard with caution
whatever divisions have been marked within our texts: be it the
paragraphs, marked by indentation, or the sentences, marked by
a running number and a stop, or the phrases, marked by a comma.
The notations are a compromise. If anyone feels the urge to in-
quire more deeply into this; problem of the interrelation between
deep structure units of content and surface structure boundary
signals he may find, thanks to our careful notation of such signals
as glottalization, emphatic stress and different pauses, ample
material for constructing an appropriate theory provided that he
knows the language very well.
The following symbols are used:
[ ] for completions of words that are not audible from the
14 Cahuuia Texts
recording but that, according to careful extrapolation, must
nevertheless have been pronounced.
[ to indicate poor recording and uncertain transcription up
to the next following word boundary*
3. 4. Representation of the Translation
Below 4,4. we discuss our reasons for giving a target-
oriented translation, not a source -oriented one. One of the
necessary consequences of our general attitude to this problem
would be to divide the English translation into paragraphs, sen-
tences, phrases, etc. , independent of the divisions we marked
in our original Cahuilla text. Instead, a compromise has been
adopted here too. Indenting is applied in the English translation
wherever a paragraph division has been recognized in the corres-
ponding words of the Cahuilla original. The length of the senten-
ces of the English translation has been brought into a near cor-
respondence with the Cahuilla sentences. Thus, the running
numbers of sentences in the original match those of the transla-
tion fairly well. Of course, this was done for those readers —
and they may constitute the majority — who are sufficiently
interested in the details of the morphemic and lexemic structure
of the Cahuilla sentences, yet unable to understand the original
without the constant help of a translation. In doing so we have
not remained true to our outspoken principles on what kind of in-
formation a translation should contain. But as things are, one
can hardly see how such a sacrifice could have been avoided.
The following symbols are used in the translation:
[ ] for additions not corresponding to any particular mor-
pheme sequence in the original, but necessary for a full under-
standing of the intended meaning of the original.
( ) for substitutions, mostly of personal names for pronouns
where the reference is not clear, e. g. *he (Mukat) . . . ' On the
problem of personal reference, see 4. 4.
Words in Songs were translated where possible, e. g. the
songs contained in Kunvaxmal. Songs containing Cahuilla words
but difficult to translate were given in the same transcription —
15
a phonemic on$ — as the original, e. g. Creation, 282, 285.
Meaningless syllables of magic chants have been transcribed
phonemically in the original; but in the English translation an
anglicizing transcription has been substituted so the English read-
er, pronouncing these sequences a l'anglaise will get an impres-
sion of their phonetic shape.
4. On the Interrelation between Text , Translation , and
Grammar
4. 1. Introduction
According to the standards set for the description of an Amer-
ican Indian Language by the exemplary works of Bloomfield and
Sapir and perpetuated by such institutions as the Survey of Cali-
fornia Indian Languages, a complete description should consist of
a grammar, a dictionary and a collection of texts with translation.
While linguists have in recent times given a considerable amount
of thought to the question of what kind of information is or should
be presented in a grammar, the analogous question about the in-
formation presented in a text representation or in a translation
has not yet been raised. We thus propose to raise, and if possible,
to answer the question: what kinds of information should be pre-
sented in a grammar, in a text, and in a translation? The answer,
we think, must come from a close scrutiny of the interrelation
between these fields of linguistics activity. Modern linguisitics,
especially generative grammar in its various models, has put
strong emphasis upon describing the linguistic competence of a
native speaker. Linguistic competence, in this view, is explicated
by the rules which generate all the grammatical sentences and
only the grammatical sentences of a language. However, there
can be no doubt as to the existence of a competence of performance.
Concrete examples will be given in section 3. The phenomena ex-
emplified must be part of the speaker's internalized knowledge of
the language. Competence of performance manifests itself in the
texts only. More precisely, it becomes apparent in the phonic
representation of a text and in the description of its content* While
we do not as yet dispose of a firmly established model and method-
ology of content description, it is safe to say that the problems of
16 Cahuilla Texts
content description become apparent in translation. Examples
of this will be found in section 4. 3. 3.
Thus, the view is held here that linguistics is not coexten-
sive with grammar making; that both the representation of the
phonic aspect of a text and its translation into one of the inter-
national languages are integral parts of the linguistic descrip-
tion of a language; and that we must remain open to the possibility
of integrating further aspects into the overall task of describing
a language*
The study of hitherto unexplored languages offers an instruc-
tive test case for these views, because neither a grammar, nor
a text collection, nor translations are given beforehand on which
to build. In the famous passage of the Cours de linguistique ge-
nerate (p. 2 3. ) , F. de Saussure raises the question of the object
of linguistics. *While other sciences operate on things given
beforehand which are then considered from different angles, we
find nothing comparable in our field. Suppose someone utters
the French word nu: a superficial observer might be tempted to
see here a concrete linguistic thing; but a more attentive exam-
ination will successively reveal three or four completely differ-
ent things depending on the point of view adopted: a sound, the
expression of a concept, the correspondent of the Latin word
nudum, etc. Thus the thing, far from preceding the point of view,
is, one might say, created by it; and who can tell us whether one
or other of these different ways of considering an object is anter-
ior or superior to the rest?' The translation is mine.
If these words are substantially true, as we think they are 3 $
we must inquire next where these different viewpoints which
'create' the object of linguistics stem from. In the following
theoretical remarks we shall distinguish between two fundamental
situations inherent in every verbal communication, each giving
rise to different sets of questions and each calling for different
sets of answers.
4, 2. Four Postulates
Introduction 1 V
4. 2. 1. Two Fundamental Situations
It vas intimated, in 4. 1. that the viewpoints which 'create*
the object of linguistics result from fundamental situations,
positions or attitudes which are inherent in speech communica-
tion. We call these situations:
(a) 'From inside 1
(b) 'From outside 1
(a) corresponds to the situation or position of the native
speaker of a language A; it corresponds to his internalized tacit
knowledge of the regularities of A. The attitude of (a) is: *I
know everything. ' The problem of describing A from the point
of view of (a) is to make the internalized tacit knowledge explicit.
Since the internalized knowledge refers to a system of rules or
regularities which has the effect of 'pairing 1 a certain sequence
of sounds with a certain content, the central problem consists of
explicating this mechanism intervening between sound and content.
(b) corresponds to the situation or position of the native
speaker of a language B when confronted with language A in an
act of communication. Such a situation may occur at any time
when people communicate verbally. The difference between A
and B may be a slight one as between closely related dialects or
as between the speech of the older and the speech of the younger
generation of a speech community, or the difference may be a
substantial one as between two unrelated languages, (b) , the
position of a native speaker of B confronted with A is the position
'from outside 1 ; he does not have at his command the full inter-
nalized knowledge of A.
It might be objected that (b) is a matter of performance,
while (a) is a matter of competence. However, we have already
postulated a competence of performance. This is tantamount to
saying that certain phenomena that were attributed to performance
and therefore remained outside the range of interest of generative
grammarians must be given the attention and the status in theory
they deserve. Among these phenomena we reckon the process of
learning and describing a foreign language. Just as there is th«e
18 Cahuilla Texts
problem of describing A from the point of view of (a) , there is
also the genuine problem of describing A from the point of view
of (b). It is the problem of all field work. This problerri, which
we may symbolize as (b) /A differs from the problem (a) /A in
that (a) but not (b) may use the full implicit knowledge about the
language he is describing in order to test his hypotheses. What
at first sight seems to be given for (b) is, by way of perception,
the acoustic speech signal and, by way of translation, the mean-
ing content. Now, it is a fact that foreign languages can be ac-
quired and that they can be described. Thus the problem (b) /A
lends itself to a solution which with progressing field work grad-
ually approaches the solution of problem (a) /A, though probably
without ever completely matching it. A problem solver of the
type (b) /A must proceed with particular care. When trying to
state the regularities of A, (b) must again and again revert to
the speech signal on the one hand and to the corresponding con-
tent on the other. Here an attitude of 6 I don't know anything*
may be superimposed on or even contradict the above described
attitude of *I know everything. *
The situation (b) /A is just as fundamental in speech commun-
ication as is (a) /A. Both must be recognized to have theoretical
status. Yet (b) /A is theoretically interesting in a different way
from (a) /A. While in (a) /A the theoretical interest is focused
upon the competence of a native speaker of A, the theoretical
interest of (b) /A centers around the paths which (b) follows to-
wards the solution of the problem of describing A as a foreign
language. Very little is known empirically and in theory about
these paths; a wide field is open here for future research.
4. 2. 2. Presence of both Situations
It seems necessary that in theory the two fundamental situa-
tions be clearly distinguished. The informed reader may already
have correlated them with the two most conflicting models of
modern linguistic description: American structuralism and trans-
formational generative grammar. In fact their paths of thought
roughly correspond to the paths taken in solving problems of types
(b) /A and (a) /A respectively. Since in type (b) /A the implicit
knowledge about the regularities intervening between sound and
Introduction 19
content is incomplete, it seems understandable to exercise cau-
tion and even to ignore part of the knowledge one actually has,
as well as to search lor discovery procedures. Since in type
(a) /A problems an implicit knowledge about the regularities may
be advocated for testing hypotheses, generative grammar empha-
sizes derivation and finds no room for discovery procedure in its
theoretical framework.
However, our thesis here is that such a sharp division, though
necessary in theory, is not found in the reality of verbal commun-
ication. What we find, instead, is that one fundamental situation
predominates over the other, but never to its complete exclusion.
A problem solver of the type (b) /A is never entirely ignorant
of the abstract underlying mechanisms and structures of A to be
described in the grammar of A. A case in point are the so-called
universals of language. While linguists can't agree over the re-
cognition of specific universals, their existence in principle seems
generally accepted.
A problem solver of the type (a) /A, though implicitly having
at his command the abstract underlying mechanisms and structures,
is given the additional faculty of partially ignoring this knowledge
at any time that he so wishes. This extra capacity of a native
speaker is a very basic part of his competence, which, among
other things, makes it possible for native speakers to become lin-
guists. The possibility of excluding from one's consideration —
however temporarily — certain aspects of linguistic competence,
is based on this very faculty. To cite an example: In the 1957
model of generative grammar future research into semantics was
postponed since it was said that semantics presupposes syntax.
The art of punning or of rhyming, which often entails a conscious
disregard of morphemic structure in order to highlight phonolog-
ical similarities, at the same time presupposes the full internal-
ized knowledge and the need as well as the possibility of disre-
garding part of that very knowledge.
4. 2. 3. Grammar , Text , and Translation
How are the two fundamental situations we postulated earlier
in this chapter, represented in a grammar, in a text and in its
translation?
20 Cahuilla
We have seen that the fundamental situations do not occur in
their pure, theoretical form, but that there is a predominance of
the one or the other without either of them being completely ab-
sent.
4. 2. 3. 1. Grammar
A grammar may be viewed, in the most general terms, as a
device that explicates the association between certain sequences
of physically observable vocal sounds emitted by a native speaker
and certain inner thoughts or ideas, called content, of that same
speaker. The grammarian tries to explain this phenomenon of a
pairing between sounds and content by postulating that one who
knows a natural language tacitly knows a finite system of rules
enabling him to produce and understand an infinite number of sen-
tences, some of which he has never previously encountered. What
needs to be explained, then, is precisely this knowledge. Clearly,
an idealization is necessary here: only a full or complete implicit
knowledge of the system of rules can guarantee the production of
all and only the grammatical sentences of a language.
Our thesis here is, that grammar predominantly corresponds
to the fundamental situation (a) with regard to A. The path the
explication must take, then is the path leading from a system of
abstract representations and of rules towards the vocal signal.
4. 2. 3. 2. Text
We first raise the question: what is a text? Astonishingly
little thinking has been done about this problem until very recently.
Some relevant contributions are now emerging from the European
scene 4 . According to our own investigations two entirely different
things are often confounded by using one and the same designation
*text':
1. *What a speaker tells/ as e. g. a story.
2. The representation of what a speaker tells e. g. a graphic
representation.
Introduction 2 1
While we have denounced (section following de Saussure) the idea
that objects of linguistic research are given beforehand, inter-
pretation 1. in fact seems to presuppose just such an object.
But how can we possibly think of *what a speaker tells, * how
can we manipulate this phenomenon in our mind? The only way
is by representation. Even the most direct replica of *what a
speaker tells, * the recording on magnetic tape, is a representa-
tion resulting froir .ne application of certain definite principles
of filtering out and thus of abstraction. A tape recording is
therefore a representation comparable to any other representa-
tion such as by phonetic symbols. Now, representations con-
stitute the answers to problems and questions raised. What are
these questions and what are the corresponding answers involved
in text representation? In other words: What kind of information
should a text representation contain? Questions, as it was said
in section 1. , originate from certain fundamental situations. The
two fundamental situations discussed earlier prove to be relevant
here, too.
A fragment of text in a language, e. g. an utterance in A,
may be advocated among other things in order to refute a hypoth-
esis concerning the grammatical structure of A. For such pur-
poses texts have been utilized both by native speakers of A and by
non-natives. If A is an aboriginal language, it has at least hither-
to been the norm that representations of texts have been elaborated
by and published for non-natives. A notable exception, which how-
ever proves the rule, is the impressive collection of first readers
and primers written by members of the Summer Institute of Lin-
guistics. Representations of this kind, which are intended to serve
the needs of indigenous people, are currently given the name of
orthography 5 . We may speak of an orthography in the case of
normalized representations where the criteria for normalization
may not uniquely be derived from the structure of the language
represented, as e. g. when the normalization is intended to facil-
itate reading for speakers of a neighboring language of trade
(Spanish in Mexico} .
Now if our text representation is not an orthography in this
sense and if it is composed by non-natives and intended for the
use of non-natives who are only in the process of mapping out the
22 Cahuilla Texts
full set of regularities of language A, our thesis regarding text
representation would be as follows: The representation embodies
regularities underlying the actual speech signal; but the repre-
sentation must be close enough to the signal to permit the non-
native linguist to follow a path leading from the signal towards
this representation and towards deeper ones which he might wish
to establish. It is a representation which artificially abstracts
from part of the knowledge a native speaker has, in particular
from part of the morpho-syntactic information. But we have
seen above that abstractions of this kind constitute an aspect of
the competence of the users of natural language and that such
artificial abstractions in themselves constitute a linguistic fact
and must be given status in a comprehensive theory. Our repre-
sentation tries to harmonize phonological and morphological in-
formation so as to make it possible to base the representation
on purely phonological rules. The rationale following from this
thesis will become understandable when we discuss concrete ex-
amples (below 4. 3. 2. ) .
4. 2. 3. 3. Translation
Neither a text nor its translation into a different language are
things that can be taken for granted. Thus, we again raise the
question: What kind of information should a translation contain?
Our thesis here is that a translation should render as faith-
fully as possible the cognitive content of the original text. All
then seems to depend on the notion of cognitivity. For us this is
an empirical question which will be answered by discussing ex-
amples in 4. 3. 3.
One of the basic controversies about translation is whether
it should be literal or free. E. Nida distinguishes between two
basic attitudes which he terms *Formal Equivalence* (F-E) and
*Dynamic Equivalence' (D-E). A translation following the F-E-
principle is basically source -oriented, a D-E-translation is
oriented towards the receiver and his response. While F-E aims
at equivalence with regard to the forms of the source language,
D-E tries to achieve equivalence with regard to the reactions of
the receiver in the target language. As Nida very aptly remarks 7 ,
an evaluation of different translations must first of all take into
introduction 2 5
possibility — which will be corroborated by empirical consider-
ations (4. 3. 3. ) — that not every kind of content information con-
tained in a text might be reproduced in the translation, the trans-
lation containing in principle the *cognitive content* of a text. If
this is accepted the question then arises as to where in the
*organon' this kind of information not reproduced in the transla-
tion should be represented.
A new section of the 'organon', not provided for in tradition-
al linguistics, might be set up. The name — pragmatics might
qualify — is of secondary importance. What we are trying to
say here is that certain facets (one might also call them features)
may combine into new kinds of problems or questions calling for
new kinds of answers. This in turn would mean that the rigidly
compartmentalizing view of the *section* (of grammar, text,
translation, as pidgeonholes for problems) should be given up in
favour of a view which admits cross -classification of problems.
Thus, a problem like the earlier mentioned puns might show one
feature pertaining to the (a) /A and one feature pertaining to the
(b) /A situation. Or, a problem might show a sociological facet
or feature and a linguistic one.
Just as there are no pre-established things or objects of
linguistic research, neither are there any pre-established domains
into which to divide the problems. This ultimately holds for lin-
guistics itself: the view of a firmly delimited field of research
should be abandoned. In practice, this has already been done,
the proliferation of such compound designations as sociolinguistics
and psycholinguistics bears witness to it.
^* ^* Empirical Considerations
4. 3. 1. Grammar
In the section on grammar (4. 2. 3. 1. ) it was stated that a
grammar must explicate the association between sound and con-
tent. We know that generative transformational grammar claims
to do this by a finite system of rules, and that the system con-
sists of a base including a categorial component plus a lexicon,
and of a set of transformational rules. The base is said to con-
tain all the information necessary for the semantic interpretation.
26 Cahuilla. Texts
However, in recent developments of transformational work
it has become more and more evident that large portions of con-
tent transmitted from speaker to hearer and belonging to a
speaker 1 s ability to produce and understand sentences have sim-
ply been omitted from treatment in the base of a grammar.
These portions comprise such information about the situation of
the speaker as his social status, his geographical and temporal
location at the time of the speech event. Various proposals have
been advanced to cover these kinds of information within the
framework of generative grammar. The procedures proposed
may be roughly divided into two classes:
(a) The information is introduced directly into the system
of generative rules. Sadoc^s Hypersentences * are of this kind;
also Fillmore ! s entailment rules 11 .
(b) The information is introduced indirectly by metalinguis-
tic statements. Proposals of this kind were advanced among
others, by McCawley 12 and Wunderlich .
It is not our intention to discuss these proposals here. Let
it suffice to say that the corresponding problems are basically
the same as those treated by us in the section on translation
(4. 3. 3. )• Incidentally, it is while working out a translation that
one becomes aware of many content features not yet accounted
for in grammar. The main problem is not whether to pidgeon-
hole them under Grammar or Translation, but to characterize
them exhaustively by pointing out their distinctive features or
facets
The only point of grammar we want to examine more closely
belongs to phonology. Here it is much easier than in morphology
and syntax to justify an abstract underlying representation and a
set of rules indicating the path leading towards the signal.
One of the salient features of our underlying representations
is the separation by space of word forms. In the utterances act-
ually heard the word boundaries are often blurred by optional
phenomena of Sandhi. Consider the following utterance in its ab-
stract representation.
Introduction 27
(1) 1 1 mutuleka hemk w e ? eqewen| |
'they were (are) rising early'
Each word is marked by a primary accent. The following
points of morphological structure need to be retained here:
mutuleka *early' shows a suffix - ka indicating direction in time
or space,
hemk w e ? eqewen *they were (are) rising 1 is a verb form and
contains a root -k w £ ? e - 'rise 1 , a subject prefix of the third
person plural and suffixes. Now, an optional contraction
may occur between the two words when the first ends in a
vowel and the second begins with a prefix showing the phonic
sequence hV.
In the case of the example we chose three phonetic realiza-
tions are possible:
(2 ) (i) [mutulekahemk w e' ? eqewen]
i. e. no contraction has taken place,
(ii) [mutuleke • mk w e* ? eqewen]
i. e. the final vowel and the initial h have been dropped,
the vowel following the h is lengthened. This is a rule
we would call Contr^.
(iii) [mutuleka * mk w e" ? eqewen].
i. e. the final vowel is lengthened, the initial h plus
vowel are dropped. This rule we would call Contra*
Loss of intervocalic h with compensatory lengthening of a
vowel is also known even where the conditions stated above do not
prevail e, g. , when the vowels preceding and following the h are
homorganic. What is troublesome in our example is that the con-
traction is optional and that there is a choice between two differ-
ent sets of sandhi rules. It has not been possible so far to spec-
ify the conditions that trigger the contraction and that determine
the choice between Contr^ and Contr^.
28 Cahuilla Texts
4. 3. 2. Text
We follow the line of thought indicated by continuing to use
the example discussed in the preceding section. The conditions
so far unspecified may pertain to various orders. They may have
to do with speed, or with social status, or even with dialect.
Evidently then, it has not yet been possible to explicate the
full internalized knowledge of a native speaker regarding these
structural connections. But one should add that this is the nor-
mal situation vis-a-vis a text in a foreign language.
Reminding ourselves of our general theses regarding text
representation we would now raise the question of what would
be the appropriate representation in the case of our example?
One might consider a narrow phonetic transcription to be a
possible solution. However, we see no reason why all the cum-
bersome phonetic detail should be presented and why the regular-
ities that may be formulated in phonological terms should be
ignored. Taking into account these regularities would certainly
make for a more perspicuous representation. More important:
a text representation is meant to be used in conjunction with a
grammar and a dictionary. Difficulties are bound to arise if the
underlying representation employed in the grammar and diction-
ary and the text representation are too far apart from one an-
othe r.
On the other hand one might choose the abstract underlying
representation for the text as well. However, in the case of our
example, several pieces of information would then have to be
dropped: information on whether, in this particular instance, a
contraction takes place or not; and if it does whether Contri or
Contr2 would be chosen. Once given up in favour of a normali-
zation, these pieces of information can no longer be recovered.
They may represent details of minor importance or they may on
the other hand represent some missing links in some important
structural connection.
Clearly both of these radical solutions to the problem of a
text representation show serious disadvantages. What seems to
be called for, then, is a compromise. Where phenomena of
Sandhi occur we would give the information which comes from the
speech signal preference over the information which comes from
Introduction
morphology. Thus, we would write without space
(3) /mutuleke *mk w e* ? eqewen/ or
(4) /mutuleka # mk w e* ? eqewen/
as the case may be. This notation roughly corresponds to a
phonemic one. Where no Sandhi phenomena occur, we see no
reason to omit from the representation information as to mor-
phological structure and boundaries of words. By introducing
once and for all the convention that space is phonologically irrele-
vant, we may write in our texts
(5) /taxliswetem hemk w e ? eqewen/
*the people were (are) rising'
which would be identical with such an abstract representation as
(6) || taxliswetem hemk w e* 9 eqewen 1 1.
4. 3. 3. Translation
We should like to present the reader with certain classes of
content phenomena for which it was difficult if not impossible to
find an English translation. More precisely, it was not possible
to find constant and principled ways of transposing them into a
different language.
The classes of phenomena have a common denominator in
that they all include reference to the speaker and to the speech
act*
4. 3. 3. 1. Deictic Adverbs
In Cahuilla, we find a class of local adverbs which show the
following contrastive semantic components:
1. direction vs. non-direction (or place) ,
2. proximal vs. medial vs. distal deixis .
30 Cahuilla Texts
Informants will unhesitatingly give the following translation
equivalents for proximal: ? fpa 'here, * ? fka 'in this direction, '
? ipax 'from here, * ? fpika 'towards here. ' For medial: ? £r|a
*there, where you are' (correlative). For distal: peqa 'there,
far away, * pxka 'to a far away place (one cannot see) \ peiqax
'from far. '
Our primary interest is in the proximal and the distal sets.
Proximal clearly involves the point of view of the speaker. Dis-
tal, unmarked in this respect, in some cases does not specify
proximity. In others it implies distance, the place without direct
connection with the speaker.
There are two striking things about these adverbs: 1. Their
frequency: every verb form in Cahuilla text may be and often is
construed with a local adverb. 2. Their use: both proximal
and distal deixis occur in dialogues as well as in narrative prose.
Whereas in a dialogue it is natural from our point of view that
the speaker 1 s location is constantly involved and that proximal
deixis is expressed, in narrative prose we find it normal to re-
frain from introducing the speaker ! s (i.e. the narrator's) point
of view and to leave proximal deixis unexpressed. What we find
in Cahuilla narrative prose, however, is an unrestricted use of
adverbs showing proximal deixis. Consider the example
(7) kelawati pe*mhivinwen me* mvukqekpaywen ta ? ,
me •mc'exinwen ? fkai
'They took up wooden sticks and hit them (the geese) ,
they killed them in this direction*
Upon closer inspection of all that is entailed in this and in similar
examples we find that the contrastive semantic features of prox-
imal vs. distal may be correlated both with the narrator's situa-
tion and with the point of view of the agent of the verb phrase to
which the adverb belongs. In other words: what coincides with
the agent's localization is proximal, what is distant from the agent
is distal. Thus, in our example: '. . . they killed them [- an action
which originated from them and went the way of their intention -]* .
The native informant, however, when asked about the meaning of
?ika in this and similar examples would unhesitatingly answer:
"The direction in which I am pointing. '
Introduction
This complexity may be illuminated by pointing to the quite
analogous use of a present tense in narrative prose in many-
languages, e. g. the Praeeens Historicum in the classical lan-
guages. The present tense, like proximal deixis, implies re-
ference to the speech event. And it, like proximal deixis in
Cahuilla, may also be correlated with the narrated event. It is
precisely this twofold reference which causes the impression of
a 'vivid or dramatic narration. '
In a translation into English or German, it is impossible to
find a principled way of transposing this double reference in prox-
imal deixis. Each local adverb would necessitate comments of
the sort just made. This, however, I would no longer call a
translation. Because of these difficulities we would propose to
leave unexpressed the double reference and, in the case of our
example, simply indicate when the action shows a direction orig-
inating either from the agent or from the narrator. Such a pro-
cedure would result in a certain loss of information for which,
apparently, the translation cannot make up. No doubt we are pre-
sented here with a feature of content structure which is controlled
by the native speaker and thus forms part of his internalized
knowledge. The question arises as to whether the grammar should
account for such features or whether an independent section should
be set up for this and similar purposes. Much more important,
however, is the task of explicitly specifying the problem in all its
facets.
4. 3. 3. 2. Personal Reference
The problems to be dealt with in this section pertain to the
identification and change of the participants in the narrated events.
In Cahuilla narrative style the participant persons must be ident-
ified by their name only at their first appearance or in disjunctions
of the type 'either A or B'. Otherwise the various persons are
normally referred to as pe* 9 'he, she, it' which functions as a pro-
noun third person and shows distal deixis. Since it can only be
differentiated in number but not in gender, a considerable uncer-
tainty arises for the non-native listener — not for the native nar-
rator — as to which person is being referred to. An English or
German translation cannot possibly imitate this unspecified way
32 Cabuilla Texts
of personal reference: explicit names or kinship terms have to
be substituted. This, however, is in certain definite ways con-
trary to the intentions of the original text, which are to leave
specific personal identification unexpressed wherever possible,
or even to employ circumlocutory expressions in order to weak-
en the directness of identification. The following expressions
provide examples of this feature:
(8) pi? tikuS 'he, Takus"
The personal name is not given directly but in an appositive con-
struction.
(9) pe"> hi^itete 9 takuS
'he, what was (is) his name, Taku¥*
Again the personal name is introduced as an apposition to the
personal pronoun plus a pro-form explicitly stating uncertainty.
The frequence of expressions like (8) and (9) in Cahuilla narra-
tive texts contrasts in a striking way with the awareness of the
natives as to which person is being referred to, an awareness
which we have tested at regular intervals in order to be sure.
Again we are presented with a very special content feature:
avoidance of personal identification in spite of factual unambiguity.
Without any doubt it is part of a native speaker's ability to con-
trol this feature. No doubt either that the problem has both a
linguistic and an extralinguistic facet. As for the extralinguistic
facet we would have to inquire into the role of naming and person-
al reference within the overall sociological code of the Cahuilla
people — an inquiry which would probably come too late in view
of the rapid disintegration of this tribe.
In any case it seems doubtful whether facets of this sort can
be accounted for in the grammar or whether a special section,
say, sociolinguistics, should be introduced. It is, however, far
more important to give an explicit statement of the different facets
of the problem.
Introduction 33
4. 3. 3. 3. Status of the Proposition
There are two contrastive expressions to assert the exis-
tence of objects:
(10) ? iyaxwen 'it exists in this way'
'is this way*
which is a verb form and may be analyzed into a demonstrative
prefix 2±T* * ne verD root - yax - 'manifestation of some quality
or manner*, a stem affix - we n - 'durative'. This expression
asserts existence in a definite quality or way, as in
(11) tamit pfka ? iyaxwenepa
'when the sun was this way, in that direction'
(indication of time)
The other existential predication is:
(12) miyaxwen 'it exists in an underterminate
way*
'it is somehow'
a verb form analyzable into an indefinite or interrogative prefix
rn-, the verb root and affix as in (10) . This expression asserts
existence in an indefinite or uncertain quality or way, as in
(13) taxmu'at piyk miyaxwen
'there is a song about it, * literally:
'a song is somehow to it'
(14) hunwet miyaxwen
'he is a bear, ' literally:
'he somehow' exists as a bear'
(15) tuku kuyil miyaxwe ? ne
'yesterday there was a funeral, ' literally:
'yesterday there was somehow a funeral. *
34 Cahuilla Texts
The literal translation, which takes indeliniteness into account,
seems odd to us, given the factual character of the asserted events.
In Cahuilla the existence of an object has to be expressed as a
qualified existence, whereby the qualification may either be demon-
strative ('in this way*) or indefinite ('somehow') . Since there is
no correlate to the English or German existential predication with-
out qualification, one might consider ignoring in translation the
indefinite prefix as the unmarked member of the opposition. This,
however, would lead to a loss of information. For not only may
existential predication be qualified as uncertain, there are othet
indications in our texts that the Cahuillas, in contradistinction to
speakers of West -European languages, express factual and observ-
able things as if they were uncertain. One such indication is the
frequency of such particles or phrases of indecision. Or doubt £s
(16) (i) 9 esan *I guess'
(ii) warn 'maybe *
(iii) ha* ki ? i 'or not'
in description of things or situations that can be seen both by the
informant and by the linguist 15 .
Thus, in Cahuilla existential predication has to have a status
either of definiteness or inde finite ness, and it is part of a native
speaker's ability to control these features appropriately. It
seems difficult, however, to find a purely linguistic explanation
for the reasons why the existence of certain objects is qualified
in a definite way whereas the existence of other objects is marked
indefinite. In other words, it is doubtful whether all the relevant
structural connections with these empirical facts may be found
within the domain of verbal behaviour. The problem in addition
to the linguistic facets outlined above may have one or more ex-
tralinguistic facets. In consequence it seems doubtful whether it
would be appropriate to account for these facts of content structure
in a grammar.
4. 4. Conclusion
As stated before, the phenomena described under 4. 3. 3. 1. -
4. 3. 3. 3. have a common denominator: They indicate the speaker 1 s
Introduction 35
situation (locational, sociological, psychological) vis-a-vis the
utterance. All these phenomena proved to be difficult if not
impossible to account for in translation.
Complying now with the task of distinguishing cognitive from
non-cognitive content (see above 4. 2. 3. 3. ) , we might state that
these phenomena constitute the class of non-cognitive features
of content. And we might further state that these and only these
may be ignored in translation, with the proviso, however, that
the information should not be dropped altogether but accounted
for in another section or compartment of the *organon\
Considering which would then be the appropriate section we
found that grammar would not qualify. One might, therefore, as
intimated before, wish to establish a new compartment to which
one might give such names as pragmatolinguistics (including
sociolinguistics, psycholinguistics and others) or simply prag-
matics. Such a new compartment might correspond to a further
fundamental situation in verbal communication which would have
to be added to the two fundamental situations (a) 'From Inside 1 ,
(b) 'From Outside* mentioned in 4. 2. 1. This further fundamental
situation would be characterized thus: (c) 'Pragmatics is involved
in every speech act* .
The proposal to recognize a separate field of pragmatics is
not new , However, we must warn all those willing to accept
such a proposal without further questioning. The findings of our
study may be summarized in the statement:
There is a basic antinomy between two views which we might call
compartmentalisation (or classification ) and cross -classification .
Compartmentalisation seems to correspond to fundamental situa-
tions in speech communication. It results in the establishment of
such domains as grammar, text, translation, pragmatics. Cross-
classification corresponds to the features or facets of problems
which we have attempted to point out in the course of our research.
It is true that there is no cross-classification without classes.
But it is also true that classes are based on qualities or features.
The view of grammar as a monolith must be abandonned.
The concept of problem features serves to explicate the interrelation
36 Cahuilla Texts
between grammar, text, translation, pragmatics, and other
possible domains.
After all that has been said before, it seems that linguistics
in the usual sense is too narrow a platform from which to discuss
the interrelation between these domains. A broader frame of
reference would be needed here. Such a frame of reference could
very well be represented by the new science which, following Ch.
S. Peirce, has been called Semiotic and which, at the present, is
in a phase of vigorous expansion. The topic, however, cannot be
treated here as it would go far beyond the scope of the present
study.
INDIANA UNIVERSITY PUBLICATIONS
LANGUAGE SCIENCE MONOGRAPHS
Editor: C. F. Voegelin
Volume 6
Hansjakob Seiler
CAHUILLA
TEXTS
WITH AN
INTRODUCTION
Copyright © 1970 by Indiana University
All rights reserved
Standard Book Number 87750-148-3
Library of Congress Catalog Card Number: 72-630298
All orders from the United States of America and from Canada should be sent to Humanities
Press, Inc., 303 Park Avenue South, New York, New York 10010. Orders from all other
countries should be sent to Mouton and Co., The Hague, The Netherlands.
Printed in the United States of America
Published by
INDIANA UNIVERSITY, BLOOMINGTON
Mouton & Co., The Hague, The Netherlands
MK