Eugene Nida’s Principles of Translation as Exemplified by Bible Translating

In terms of the length of tradition, the volume of work, and a variety of problems, Bible translating is distinctive. Beginning with the translation of the Hebrew Old Testament into Greek in the second and third centuries B.C. and continuing down to the present time, the Scriptures have been translated, at least in part, into 1,109 languages, of which 210 possess the entire Bible and 271 more the New Testament. This means that the major part of the Christian Scriptures exist in the languages of at least 95 percent of the world’s population.

Moreover, most of this work has been accomplished in relatively recent times. By the time of the invention of printing, approximately 500 years ago, only 33 languages had anything of the Bible, and even by the beginning of the nineteenth century, only 71 languages possessed anything of the Scriptures. However, within the nineteenth century, more than 400 languages received something of the Scriptures, and during the first half of the twentieth century, some part of the Bible was translated into approximately 500 more languages and dialects. At present, the volume of translation and revision is of such magnitude that within the next twenty-five years as much will be published as within the entire nineteenth century, for more than a thousand persons are giving all or a major part of their time to the translation and revision of the Bible in various parts of the world.

The unparalleled range of Bible translating, including as it does not only all the major languages of the world but hundreds of “primitive” tongues, provides a wealth of data and background of experiencing the fundamental problems of communication that constitute the basis of the following article.

Practical Nature of Problems in Bible Translating

Whereas for some people translating may be primarily a matter of theoretical interest, the Bible translator must face up to certain immediate problems. For example, if he attempts to translate literally the expression “he beat his breast” (speaking of the repentant Publican,12 Eugene A. Nida Luke 18:13), he may discover that, as in the Chokwe language of Central Africa, this phrase actually means “to congratulate oneself” (the equivalent of our “pat himself on the back“). In some instances, it is necessary to say “to club one’s head.”

It is assumed by many people that the repetition of a word will make the meaning more emphatic, but this is not always the case. For example, in Hiligaynon (and a number of other Philippine languages), the very opposite is true. Accordingly, one cannot translate literally “Truly, truly, I say to you,” for to say “truly, truly” in Hiligaynon would really mean “perhaps,” while saying “truly” once is actually the Biblical equivalent.

Quite without knowing the reasons, we usually insist that, in rendering in another language a sentence such as “he went to town,” one must use an active form of the verb meaning “to go.” However, in many of the Nilotic languages of Sudan it would be much more acceptable to say, “the town was gone to by him.”

In still other instances one encounters what is regarded by some as a completely distorted orientation of experience. For example, in the Bolivian Quechua language, it is quite possible to speak of the future, even as it is in any language, but one speaks of the future as “behind oneself” and the past as “ahead of one.” When pressed for an explanation of such an expression, Quechuas have insisted that because one can see “in the mind” what has happened such events must be “in front of one,” and that since one cannot “see” the future such events must be “behind one.” Such a perspective of the past and the future is every bit as meaningful as our own, and it can certainly not be condemned as distorted. It is simply different from ours.

Accordingly, in such areas as (1) behavior as described by language (e.g., “beating the breast”), (2) semantic patterns (e.g., repetition of constituents), (3) grammatical constructions (e.g., active vs. passive), or (4) idiomatic descriptions of “perspectives,” the Bible translator is faced with acute problems demanding answers. He knows full well that reproducing the precise corresponding word may utterly distort the meaning. Accordingly, he has been obliged to adjust the verbal form of the translation to the requirements of the communicative process.

Underlying Principles

Though in many instances the principles underlying Bible translating are only partially recognized or formulated by those engaged in such work, nevertheless the results of any accurate translating reveal the following basic principles:

  1. Language consists of a systematically organized set of oral-aural symbols. By oral-aural we are simply emphasizing the fact that such
    symbols not only are uttered by the vocal apparatus of the speaker but are also received and interpreted by the listener. The writing system of any language is a dependent symbolic system and only imperfectly reflects the “spoken-heard” form of language.
  2. Associations between symbols and referents are essentially arbitrary. Even onomatopoetic forms bear only a “culturally conditioned” resemblance to the sounds which they are designed to imitate. For example, the equivalent of our tramp-tramp is kú ha in Luvale, a Bantu language of Central Africa, and mingòdongòdona in Malagasy.
  3. The segmentation of experience by speech symbols is essentially arbitrary. The different sets of words for color in various languages are perhaps the best ready evidence for such essential arbitrariness. For example, in a high percentage of African languages there are only three “color words,” corresponding to our white, black, and red, which nevertheless divide up the entire spectrum. In the Tarahumara language of Mexico, there are five basic color words, and here “blue” and “green” are subsumed under a single term. The comparison of related sets of words in any field of experience — kinship terms, body parts, or classification of plants — reveals the same essentially arbitrary type of segmentation. Since, therefore, no two languages segment experience in the same way, this means that there can never be a word-for-word type of correspondence which is fully meaningful or accurate.
  4. No two languages exhibit identical systems of organizing symbols into meaningful expressions. In all grammatical features, that is, order of words, types of dependencies, markers of such dependency relationships, and so on, each language exhibits a distinctive system.

The basic principles of translation mean that no translation in a receptor language can be the exact equivalent of the model in the source language. That is to say, all types of translation involve (1) loss of information, (2) addition of information, and/or (3) skewing of information. To understand clearly the manner in which such “distortion” takes place we must examine the ethnolinguistic design of communication.

Ethnolinguistic Design of Communication

By adopting the simpler components of the communication process and relating these to the entire communicative context, we may
construct an ethnolinguistic design of communication as shown in Figure 1.

Figure 1: Nida’s ethnolinguistic design.

In the diagram of Figure 1 S stands for the source (the speaker as source and encoder). M is the message as expressed in accordance with the particular structure (the inner square in this instance) of the language. The message may include anything from a single word to an entire utterance. R is the receptor (including decoder and receiver), and the outer square (designated by C) represents the cultural context as a whole, of which the message (as a part of the language) is itself apart and a model (compare the similarity of shapes).

It is quite impossible to deal with any language as a linguistic signal without recognizing immediately its essential relationship to the cultural context as a whole. For example, in Hebrew, the root * brk is used in the meaning of “to bless” and “to curse.” Such meanings would only be applicable in a culture in which words in certain socio-religious contexts were regarded as capable of either blessing or cursing, depending upon the purpose of the source. Similarly * qds, which is generally used in the sense of “holy,” may also designate
a temple prostitute, an association which would be impossible within our own culture, but entirely meaningful in a society which was well-acquainted with fertility cults.

This emphasis upon the relationship of M to C must not, however, constitute an excuse for unwarranted etymologizing, in which meanings are read into words from historically prior usages, for example, treating Greek ekklesia “assembly” or “church” as really meaning “called out ones” (a contention of some Bible interpreters) because of an earlier use of the compound word.

Despite the recognition of the close connection between the M and C (that is, between the realities symbolized by the inner and outer squares), we must at the same time recognize the fact that every S (source) and every R (receptor) is a different individual in accordance with his background and is hence somewhat diverse in the use and understanding of M (the message). If we may describe each person’s encoding-decoding mechanism as a kind of linguistic grid based upon the totality of his previous language experience, we must admit that each grid is different in at least some slight degree. This does not make communication impossible, but it removes the possibility of absolute equivalence and opens the way for different understandings of the same message.

In the communicative process, however, S and R generally recognize these matters of difference and tend to adjust their respective grids so as to communicate more effectively. For example, a speaker adjusts himself to his audience (if he wishes to communicate with any degree of effectiveness) and the audience, in turn, makes allowances for the background of the speaker. Furthermore, each participant in the S-M-R process is aware of such adjustments and tends to make reciprocal compensation so as to comprehend more fully and correctly.1 Communication is thus essentially a two-way process, even though one person might be doing all the speaking.

One of the essential tasks of the Bible translator is to reconstruct the communicative process as evidenced in the written record of the Bible. In other words, he must engage in what is traditionally called exegesis, but not hermeneutics, which is the interpretation of a passage in terms of its relevance to the present-day world, not to the Biblical culture.

One interesting problem in exegesis which may be treated by the method of reconstructing the communicative process is the formal differences between the phrases “kingdom of God” (used exclusively in the Gospel of Luke) and “kingdom of heaven” (used in most contexts in the Gospel of Matthew). Most Biblical scholars have regarded these two phrases as essentially equivalent, but there are some persons who insist that they refer to two different “dispensations.” The answer to such a problem consists in reconstructing the facts of the communication: the Jewish taboo avoidance of Yahweh (and by extension other terms referring to deity), the substitution of words such as “heaven,” “power,” and “majesty” for Yahweh, the Jewish background of the writer of the Gospel of Matthew, the evident Jewish audience to which the Gospel of Matthew is directed, the Greek background of Luke, the Greco-Roman audience to which the Gospel of Luke was directed, and the complete lack of any substitution device (such as “heaven” for “God”) on the part of the Greco-Roman community. These factors in the communication process when considered in the light of the total cultural context make the identification of the two phrases entirely justified.

Two-Language Model of Communication

Up to the present time, we have been discussing the translator’s task in terms of the Biblical languages, but assuming, for the sake of greater simplicity of statement, that the translator was a part of the Biblical culture. This, of course, is not true, for though he may be well acquainted with numerous aspects of this culture, he is not, nor can he ever be, anything like a fully participating member. Not only can the culture not be fully described, but it can most certainly not be reproduced — despite Alley Oop’s time-machine experiences.

The fact that English (the language which we shall, for our present purposes, assume as the language of the translator) is the means by which information concerning the Biblical culture is directly or indirectly gathered, e.g., through commentaries, dictionaries, and learned journals, is described diagrammatically in Figure 2.

FIGURE 2

In this diagram, the squares represent the Biblical language (for the sake of our diagram it makes no essential difference whether we are speaking of Greek, Hebrew, or Aramaic) and the triangles represent the “equivalent” communication in English. The subscript numerals help to identify the different components in these parallel instances of communication. A translator of the New Testament into English assumes the position of R1, even though he can only approximate the role of a New Testament receptor. At the same time, this translator becomes S2, in that he reproduces M1 as M2, so that R1 may respond in ways essentially similar to those in which the original R1 responded.

Where there is a time gap between C1 and C2 the translator (S2) can only be a kind of proxy R1. However, a bilingual translator who participates fully in two linguistic communities may fulfill a dual role by being quite validly both R1 and S2.

Figure 2 serves also to emphasize two significant factors: (1) the essential differences in the form between M1 and M2, and (2) the relationship of M1 and M2 to their respective cultural contexts. Of course, the actual situation is not as simple as the diagram would imply, for nothing so complex as a language-culture relationship can possibly be reduced to a few lines. However, the differences are present and real and can be noted in all phases of the communicative procedure. A few of these differences will enable one to understand more fully certain of the broader implications of what we are only able to hint at here.

Though as English-speaking people we employ a language which is relatively closely related to Greek (certainly in comparison with the differences between English and Hottentot), there are numerous basic differences. In the meanings of words, for example, we have relatively few close correspondences. We use love to translate certain aspects of the meanings of at least four different Greek words: aga paô, phileô, stergô, and eraô, but these words also correspond to such English meanings as “to like,” “to appreciate the value of,” “to be friendly with,” “to have affection for,” and “to have a passion for.” Even a first-year Greek student will give the meaning of logos as “word,” but the Liddell and Scott dictionary lists more than seventy different meanings — and these do not do full justice to the specialized Biblical usage. However, Greek also has two other words, epos, and rhêma, which are likewise translated as “word” in any context.

The incommensurability between Greek and English is quite evident in the differences between tense and aspect, a problem which gives constant difficulty to a translator of the New Testament. This problem is made all the more acute by the fact that the Hebrew of the Old Testament employs a tense-aspect system which is quite different from that of the Greek, but which is often reflected in the distinctive Semitic coloring of many New Testament usages.

In the matter of arrangement of words, especially in the marking of a long series of dependent phrases and clauses, the English language simply does not have the structural potentialities of Greek. Accordingly, a stretch of speech which may be a perfectly good Greek sentence (consisting, for example, of verses 3-14 of Ephesians 1) can only be rendered intelligible by several sentences in English.

Whether, then, in terms of the meanings of words or idioms (“heap coals of fire on his head,” “bowels of mercy,” or “the reins and the heart“) or of the grammatical categories or arrangements of words, M1 differs from M2. However, this is not the whole story, for most Bible translators are faced not with a two-language but a three-language communication problem.

Three-Language Diagram of Communication

By means of one’s own language — which in the case of English bears a close cognate relationship to Greek and reflects a considerable historical connection with the Biblical culture, even as Western culture took over much from the Judaeo-Greco-Roman world –one not only explores the Biblical languages but in large measure tends to mediate these data in communicating into another language. Accordingly, we may diagram this process (Figure 3).

FIGURE 3

Of course, there are a number of translators who translate “directly from the original languages,” but even then a high percentage of their responses to the forms of the original languages tend to be colored by the medium of study and analysis, namely, their own mother tongue. Their task, however, is to communicate the M1 in terms of M3, with the least possible skewing as the result of M2. The problem is made more difficult in most instances by virtue of the fact that most languages don’t have any historical connection with the Biblical languages, either by being members of the same language family or because of historical and cultural associations.

However, there is one interesting fact, namely, that the so-called Biblical culture exhibits far more similarities with more other cultures than perhaps any other one culture in the history of civilization. This is not strange, if one takes into consideration the strategic location of this culture in the Middle East, at the “crossroads of the world” and at a point from which radiated so many cultural influences. This fact makes the Bible so much more “translatable” in the many diverse cultures of the world than most books coming out of our own contemporary Western culture. This essential similarity to the cultures of so many peoples helps to explain something of the Bible’s wide appeal.

Definition of Translating

A definition of translating will inevitably depend in very large measure upon the purpose to be accomplished by the translation in question. However, since in Bible translating the purpose is not to communicate certain esoteric information about a different culture, but to so communicate that R3 may be able to respond to M3 in ways substantially similar to those in which R1 responded to M1, a definition of translating which is in accord with the best traditions of Biblical scholarship could be stated as follows: “Translating consists in producing in the receptor language the closest natural equivalent to the message of the source language, first in meaning and secondly
in style.”

This type of definition recognizes the lack of any absolute correspondence, but it does point up the importance of finding the closest equivalence. By “natural” we mean that the equivalent forms should not be “foreign” either in form (except of course for such inevitable matters as proper names) or meaning. That is to say, a good translation should not reveal its nonnative source.

It is recognized that equivalence in both meaning and style cannot always be retained — in the acrostic poems of the Old Testament, to cite an extreme example. When, therefore, one must be abandoned for the sake of the other, the meaning must have priority over the stylistic forms.

Differences of Formal Structure

In comparing the form of the Biblical message (M1) with the corresponding form that must be employed in any other language (Mx), we are immediately impressed with the marked formal differences. We cannot, however, consider all these contrasts. Nevertheless, a brief statement of such problems as diversities in (a) word classes, (b) grammatical categories, and (c) arrangements of words can be illustrative of the basic principles involved in determining what is the “closest natural equivalence” in any given situation.

Word Classes

There is a great deal of difference between languages in respect to the word classes that are used to express certain ideas, for so often what is a noun in Greek must be rendered as a verb in other languages, and what is a pronoun in Greek or Hebrew frequently must become a noun in another language. Furthermore, adjectives in Greek or Hebrew are often verb-like words in other languages. Nevertheless, behind this apparent wide discrepancy in the word classes of various languages, there are some astonishing similarities. In the first place, most languages described to date have been found to have “object words” (usually treated as noun-like words), “event words” (generally designated as verb-like), and at least some other classes, often pronouns, adjectives, and/or relational particles. What is, therefore, more significant than the apparent differences between Greek and other languages (such differences are much more evident in New Testament translating than in the Old Testament) is a fundamental agreement between languages as to classes commonly called nouns and verbs.

What we designate as noun-like words and verb-like words are predominantly those which are (1) “object words” with more or less fixed figures or forms, tree, stick, hill, river, grass, rope, stone, sun, moon, star, canoe, dog, cat, head, foot, and (2) “event words,” run, walk, jump, swim, see, hear, fight, hit, talk, make, and fly. It is possible that Gestalt psychology can provide certain important clues as to the reasons for this basic dichotomy in languages, though it is recognized that in many languages there is considerable overlapping of classes and shifting of terms from one class to another. The well-defined figure, as compared with the ground (to use Gestalt terminology), could provide us with the core of noun-like words (the so-called “object words”). The less well-defined figures representing movement, becoming, passing, or “eventing” would then be represented by the “event words,” namely, the verbs. Certain characteristics held in common by various “object words,” for example, red, yellow, true, good, kind, one, and two, would provide the abstracts generally designated as adjectives, and those designating common features of events, fast, suddenly, slowly, once, and twice, for instance, would correspond to adverbs, though in this there is also considerable overlapping and shifting of class membership. In addition to the word classes designating objects, events, and abstractions, there are the relationals, which describe relations between objects or between events. If such words are used primarily as relationals between objects, we call them preposition-like words, and if they indicate relations between events, they are generally classed as conjunctions, but here again, there is a great deal of overlapping and shifting from one class to another.

The preceding paragraphs must not be interpreted as a defense of the Indo-European word-class structure, nor of the fatal error of descriptive methodology in defining a noun as “the name of a person, place, or thing.” Furthermore, we are not suggesting that these semantically important classes represent any inevitable direction of development for any language. In the Mayan languages, for example, the equivalents of English adjectives are for the most part a formal subclass of verbs, and the prepositions and conjunctions are predominantly noun-like words, though of a very restricted class. In Tarahumara, certain object words (as judged in terms of their present semantic values) are certainly derived from event words, for example, paciki “an ear of corn” (from paci “to grow ears of corn”) and remeke “tortillas” (from reme “to make tortillas”). Nevertheless, despite such divergencies, there is in most languages a sizable core of words that reflect distinctions explicable in terms of Gestalt psychology. Moreover, whether as major or minor classes, languages do tend to have four principal groups: object words (roughly equivalent to nouns), event words (roughly equivalent to verbs), abstracts (modifiers of object and event words), and relationals (roughly equivalent to prepositions and conjunctions in the Indo-European
languages).

For the Bible translator the most serious problem relating to word classes is created by the fact that in Greek, and for that matter in most Indo-European languages, there is a marked tendency to use event words without reference to the objects or person that may participate in such events. For example, in Mark 1:4 there is the clause “John preached the baptism of repentance unto the forgiveness of sins.” All the nouns except John are essentially event words, but the participants in the events are not made explicit, and the relationships between the events are very ambiguously indicated. When, as in many languages, this type of expression must be translated not by a series of nouns but by verbs, the problem is difficult; for not only must the participants be explicitly indicated (as required by verb constructions in question), but the relationships between the events
must be more explicitly stated. This means that such an expression in many languages must be rendered as “John preached that the people should repent and be baptized so that God would forgive the evil which they had done.”

Similarly, it is quite impossible to say in many languages, “God is love.” The word indicating “love” is essentially an event word, and it cannot be combined as a kind of predicate complement to a subject by means of a copulative verb. In other words, “love” cannot exist apart from participants. One cannot say, therefore, “God is love” but simply that “God loves.” This is, of course, essentially what the Biblical passage means, not that God is to be equated with love, for the expression “God is love” can not be inverted into “Love is God.”

Grammatical Categories

When a language possesses certain categories that are not in Greek or Hebrew, the question arises as to whether the translation should conform to the categories of the receptor language. If such categories are obligatory there is really no alternative, unless one wishes to produce a translation that is grammatically incorrect. However, the problem is not quite so simple, for there are two types of factors: (1) the nonexistent, ambiguous, obscure, implicit, or explicit nature of the information in the source language, and (2) the obligatory or optional character of the category in the receptor language.

The following outline indicates those types of situations in the source and receptor languages which give rise to the most common problems of equivalence:

  • Instances in which M1 lacks information which is obligatory in M2. For example, in Matthew 4:1 there is no information available from the New Testament record as to whether Jesus had ever visited Capernaum prior to his trip recorded at this point. When, as in the Villa Alta dialect of Zapotec, spoken in southern Mexico, it is obligatory to distinguish between actions that occur for the first time with particular participants and those which are repetitious, one must make a decision, despite the lack of data in the source language. Since there is a greater likelihood that Jesus would have visited nearby Capernaum than that he would not have done so, the translation into Villa Alta Zapotec reflects this probability, and there is accordingly a distinct increase in “information” in the translation. When, however, such information is purely optional in a receptor language, it is of course not introduced.
  • Instances in which information which is obligatory in Mx is obscure in M1. The status of Jesus as a rabbi was well recognized by his friends and followers but was openly challenged by others. If, accordingly, we must apply to the Gospel accounts the categories of an honorific system (such as are common in the languages of Southeast Asia we cannot always be sure precisely what would be the relative social position of Jesus and those who would speak to and of him. Though considerable information is given, there is also real obscurity at many points. If, however, the receptor language requires honorific indicators, they must be added (with at least a partial increase in information).
  • Instances in which information which is obligatory in Mx is ambiguous in M1. Though ambiguities also involve a degree of obscurity, they are different from simple obscurities in that either alternative seems to have almost equal validity. For example, in John 4:12 the Samaritan woman speaks to Jesus of “our father Jacob, who
    gave us the well.” If we apply to this statement the inclusive-exclusive first-person plural dichotomy, which occurs in many languages, we can argue almost equally well for the inclusive form (assuming that the woman would be willing to admit that the Jews were also descended from Jacob) or the exclusive form (reflecting something
    more of the traditional hostility between the Samaritans and the Jews and the evident contrast mentioned in verse 20 of the same chapter). When the inclusive-exclusive distinction is obligatory in the receptor language, the translator must make a decision, and regardless of the results, there will be at least a partial increase of information. When, however, the receptor language allows such information to be optional, then the translator should retain the ambiguity of the original.
  • Instances in which information which must be made explicit in Mx is only implicit in M1. When information is implicit in the source language context but must be made explicit in the receptor language, there is actually no gain in the information carried by the message. It is merely carried in a different way — explicitly rather than implicitly. For example, in John 4:20, when the Samaritan woman is reported as saying, “Our fathers worshiped on this
    mountain; and you say that in Jerusalem is the place where men ought to worship,” there is no possible doubt as to the exclusive use of “our.” However, this fact is implicitly given, not explicitly so. In many instances, however, what is quite implicitly understood in one language is not so understood in another, especially in those instances where the cultural context is very different. For example, a literal translation (one which translates only the strictly explicit features) of Hebrews 7:3, “He is without father or mother or genealogy, and has neither beginning of days nor end of life…” is likely to be understood in many languages as implying that Melchizedek was a theophany, rather than simply a person for whom there is no record of human descent. Accordingly, to avoid serious misunderstanding it is often necessary to make explicit in the receptor language (even on an optional and nonobligatory basis) what is only implicit in the source language.
  • Instances in which information which is explicit in M1 must be differently treated in Mx. Explicit information in the source language should be communicated in the receptor language. There are, however, two exceptions to this general rule. In the first place, the receptor language may not have a corresponding method of indicating such information. For example, in the Greek verb system, there are numerous subtle distinctions of aspect which cannot be translated into English without very heavy circumlocutions, which in the end tend to make the aspectual distinctions far more explicit than they were in the source language. Such translations involve a partial increase in information by virtue of their emphasis. In the second place, when the indication of such information is optional in the receptor language, the frequency of occurrence of information of this type may be quite different from what it is in the source language. For example, in Greek and Hebrew numbers and tense are indicated repeatedly, while in many languages numbers and tense may be indicated once within a context, but left implicit throughout the rest of the passage in question. It is necessary that a translation indicate such optional factors with a frequency which is comparable with what would normally occur, or the translation becomes unnatural since the patterns of “redundancy” have been altered.

This outline of criteria for the addition or omission of information is applicable not only to the immediate problem of grammatical categories but to any and all types of mensurability between the source and receptor languages.

Arrangements of Words

The same principles elaborated in the preceding section with regard to corresponding categories also apply in matters of arrangements of words, whether of the order of words or of the number and types of dependencies. Of the numerous problems involved in grammatical arrangements of words, we can only touch briefly upon hypotactic and paratactic constructions. A language with a heavy hypotactic structure (e.g., Greek) simply makes explicit a number of relationships that are left implicit in a language that employs a paratactic type of structure (e.g., Hebrew). Unfortunately, there is a tendency to think that the hypotactic structure is fundamentally superior and that accordingly in translating into a language which has an essentially paratactic structure one should introduce (for example by overworking potential hypotactic patterns and by creating new grammatical forms or arrangements) the same number and types of hypotactic constructions as one finds in Greek. Such a procedure is quite unwarranted, for one should permit to be left implicit in the receptor language that is explicit in the source language if the receptor language in question would normally employ an implicit type of structure. The breaking up of long, involved sentences and the omission of corresponding conjunctions (provided such processes are carried out in conformity with the requirements of the receptor language) do not actually result in any loss of information. It simply means that the information is carried implicitly, rather than explicitly.

Hierarchy of Semantic Constituents

Despite our recognition of the fact that there are no complete synonyms, that is to say, words which may substitute for each other in all possible positions of occurrence, nevertheless, we do recognize that some words are substantially identical with others in the sense that they may be substituted for each other without any appreciable loss or change in meaning within a particular discourse. This is, of course, the experience of everyone who attempts to write without dull repetition of the same words. Not infrequently we need to mention the same referent, but stylistic considerations make it necessary for us to employ some other term that will serve the purpose. A brief examination of this process soon reveals that some words substitute for many words (words such as thing, matter, object, feature, apparatus, this, he, they, go, come, and move have a wide range of substitution), while other words may substitute for very few words (raccoon, elephant, thimble, equator, seismograph, crawl, kiss, and assassinate). If we group such words into related series and classify them on the basis of their range of substitution, we soon discover a series of hierarchies, ranging from the most concrete, “low-level” vocabulary at the base (with words having the greatest specificity), and the most generic, “high-level” vocabulary at the top (with words possessing the greatest degree of generality).

For the translator, this factor of hierarchical series of concrete-generic vocabulary poses special problems, for though languages exhibit considerable agreement as to the segmentation of experience exhibited by the concrete vocabulary (for such segmentation is dependent largely upon “figure”-“ground” contrasts which are more or less well outlined, in terms of Gestalt psychology), the generic vocabulary, which is dependent upon the recognition of common features, is much more subject to differences of interpretation. Accordingly, it is much easier for the Bible translator to translate the Book of the Revelation, which is filled with symbols, of which the meaning is obscure though the language is specific and concrete, than the Gospel of John, of which the meaning is more evident but the language of a higher hierarchical level

What makes such high-level generic vocabulary difficult to translate is not the fact that receptor languages lack such vocabulary, but that the generic vocabulary which does exist does not parallel the generic vocabulary of the Bible.

Unfortunately, there are two erroneous (and at the same time contradictory) impressions about so-called primitive languages. One often hears, on the one hand, that a language exhibits a primitive character since the language does not have any generic vocabulary, but only specific terms. On the other hand, people not infrequently lament the fact that a “primitive” language is inadequate as a means of communication because the words in question cover too wide an area of meaning, as for example in Anuak, a language of Sudan, in which the same word may designate anything made of metal, from a needle to an airplane. The actual situation that one finds in languages is not the real absence of generic vocabulary, but its occurrence on different levels, and with difficult subpatterns of substitution. For example, in Bulu, spoken in Cameroun, there are at least twenty
five terms for different kinds of baskets but no specific generic term which includes just baskets and nothing else. However, one can refer to such objects by words that would have a higher-level value than our word basket, namely, the Bulu equivalent of “thing,” “object,” or “it.” On the other hand, there are not only many different specific words for fruits but a generic term for fruits as a whole, on a level which more or less corresponds with our term. In Kaka, a related language in the eastern part of the Cameroun, there are two generic terms for fruits, one which includes bananas and pineapples, and another which includes all other kinds of fruits (in terms of our meaning of fruit), plus testicles, glands, hearts, kidneys, eyeballs, soccer balls, pills, and the seed of any fruit or plant.

Analytical studies of semantic problems in so-called primitive languages reveal that the general proportion of specific to generic vocabulary is not appreciably different from what it is in the language of so-called civilized societies. The reason for the false impressions about specific and generic vocabulary is that people have wrongly
expected generic vocabulary in various languages to exhibit the same degree of correspondence which they have observed in the study of specific vocabulary. Such is simply not the case, nor should one expect this to be so, since specific objects provide a much surer observable base of segmentation than the classification of objects,
events, abstracts, and relations, on the basis of shared or unshared features. In other words, the more one depends upon the factors of human “judgment” rather than responses to more or less immediate perception, the greater will be the tendency to diversity.

Areas of Meaning and Amount of Information

The wider the area of the meaning of a word (in terms of the wider segment of experience covered by a term) the greater is the likelihood of its statistical frequency of occurrence. This greater statistical frequency means that it tends to have a higher predictability of occurrence and hence greater redundancy. The greater the redundancy the less the information that is actually carried by the unit in question. This means that a translation made up primarily of words with wide areas of meaning does not carry the load of information which is often presumed.

There is, of course, another factor, namely, the transitional probabilities. If, for example, words with wide areas of meaning and hence greater frequency of occurrence in the language occur in unusual combinations and hence have low transitional probabilities in the particular context in question, the signal consisting of these words may still carry considerable information. Nevertheless, a translation made into any artificially restricted vocabulary will inevitably be one that carries less information than the original, unless extensive circumlocutions are employed and the meaning is thus “padded out.”

There is a tendency for translators to overwork “good terms.” They find certain expressions which may be used in a wide range of situations and hence employ them as frequently as possible. The result is often a marked rise of frequency, in contrast with normal usage, and the resultant loss in information, because of their predictability
within the Biblical context. In an analogous manner, translators often feel compelled to translate everything in the source language, to the point of employing corresponding expressions in the receptor language with an unnatural frequency. For example, in Greek, almost all sentences begin with a connective, and the result is that the connectives have relatively less meaning than the corresponding connectives in English, which occur with much less frequency. If one translates all the Greek connectives, the result is actually over translating, for the Greek words (with proportionately less meaning) are translated by corresponding English connectives (with proportionately more meaning). At the same time, while the occurrence of connectives with almost every sentence is a mark of good style in Greek, this is certainly not the case in English. This problem becomes even more acute in a language that is predominantly paratactic in structure.

Endocentric and Exocentric Structures

In the same way that there are endocentric and exocentric constructions on a formal level, there are corresponding structures on a semantic level. For example, it is quite impossible to determine the meaning of “to heap coals of fire on one’s head” by knowing the semantic distributions (types of discourse in which such words may be used) of all the component parts. The meaning of this idiom can be determined only by knowing the distribution of the unit as a whole. Accordingly, we regard it as a semantically exocentric expression. Since, however, the majority of expressions in any language are semantically endocentric, not exocentric, those who interpret the source language idioms as rendered in a receptor language are more likely than not to understand the expressions as endocentric rather than as exocentric (unless there are some special markers which provide the clues). That is the reason why, for example, in some of the languages of Congo this expression “heap coals of fire on one’s head” was regarded as an excellent new means of torturing people to death, not a means of making them ashamed by being so good to them.

The problem of endocentric interpretation of exocentric expressions can, however, be overcome in part by certain markers. For example, many of the metaphors of the Scriptures — “I am the bread of life,” “I am the door,” “a camel through a needle’s eye” — can be properly understood if they are made into similes — “I am like the bread which gives life,” or “I am like a door.” By the introduction of the equivalent of “like” the receptor is alerted to the fact that this is a kind of exocentric expression involving a “nonnormal” extension of meaning.

Similarly, the context may serve as a guide to interpretation. For example, idioms occurring in a poetical context will be more readily understood in their proper exocentric values, since the total context provides the clue to their correct interpretation.

Relationship of Linguistic Form to Semantic Function

In attempting to discover the closest natural equivalent, whether of meaning or style, one is always faced with the difficulty of finding corresponding forms with analogous semantic functions. On the level of the meaning of words in terms of their referents and their function in the cultural context (space does not permit us to deal with the parallel problems of corresponding styles), one is faced with the following types of situations:

  1. The nonexistence of a term (and its corresponding referent) in the receptor language, but with an equivalent function being performed by another referent. For example, in some languages there is no word for “snow,” for such a phenomenon is outside the realm of the people’s experience. However, the widely used equivalent of the phrase “white as snow” is “white as egret feathers.” Accordingly, in a translation, this different referent with the corresponding function may be introduced. On the other hand, if “white as egret feathers” is not a regular expression for the meaning of very white, then the introduction of “egret feathers” is not an equivalent of “snow,” and it would be more accurate to translate simply as “very, very white.” The equivalence of the two expressions “white as snow” and “white as egret feathers” is not primarily a matter of the whiteness of the respective referents, but the recognition of this fact in the traditional use of referents in both the source and the receptor languages, respectively.
  2. The existence of the referent in the receptor language, but with a different function from what it has in the source language. This means, for example, that “heart” in Greek must often be rendered by “liver,” as in the Kabba-Laka language of French Equatorial Africa, by “abdomen,” as in Conob, a Mayan language of Guatemala, and by “throat,” as in some contexts in Marshallese, a language of the South Pacific. In languages in which “gall” stands for wisdom and a “hard heart” is a symbol of courage, the Bible translator is obliged to make certain adaptations or cause serious misunderstanding.

    In some circumstances, however, the referent in the source language is such an integral part of the entire communication that it must be retained and the distinctive functions explained in footnotes. This is true, for example, of such Biblical terms as “sheep,” “sacrifices,” and “temple.”
  3. The nonexistence of the referent in the receptor language and no other referent with a parallel function. In such circumstances, the translator is obliged to borrow foreign words (with or without classifiers) or employ descriptive phrases. For example, he may borrow the names of precious stones, amethyst, ruby, pearl, or the names of classes of people, Pharisees, and Sadducees. If he adds a classifier, with resultant expressions such as “valuable stone called amethyst” and “sect called Sadducees,” he can do a good deal to compensate for the lack of correspondence between the receptor and the source language. By employing descriptive phrases, he may, for example, translate “phylacteries” as “little leather bundles having holy words written inside” (as has been done in the Navajo translation).

Within the brief scope of this essay it has been impossible to give adequate consideration to a number of significant matters: (1) stylistic parallels, a study for which certain special methods and techniques are required, (2) the influence of a translation of the Bible upon the meanings of words (that is, the important factor of the Christianization of vocabulary,” with a clear recognition of the limitations of such a process), and (3) the precise manner in which new developments in information theory, and in the broader field of cybernetics, are integrally related to Bible translating; though anyone in these fields of study will appreciate the degree to which the above analysis is dependent upon these relatively new disciplines.

In summary, however, it is essential that we point out that in Bible translating, as in almost all fields of translating, the most frequent mistakes result from a failure to make adequate syntactic adjustments in the transference of a message from one language to another. Quite sat is factory equivalents for all the words and even the idioms may have been found, but a person’s oversight or inability to rearrange the semantic units in accordance with the different syntactic structures immediately stamps a translation as being “foreign” and unnatural. These most numerous errors are not, however, the most serious, for though they may be wearisome and frustrating, they do not usually result in the serious misunderstandings which arise because of a lack of cultural adjustments.

When there are inadequate equivalents in the formal patterning of sentences (i.e., mistakes in syntax), we generally recognize such faults as once and either excuse them, or at least are able to discount them in trying to ascertain the meaning. Mistakes in cultural equivalence, however, do not carry with them such obvious clues, and hence the lack of agreement is not understood nor the source of the error detectable from the text itself.

Though it is fully recognized that absolute communication is quite impossible, nevertheless, very close approximations to the standard of natural equivalence may be obtained, but only if the translations reflect a high degree of sensitivity to different syntactic structures and result from clear insights into cultural diversities.