Concept of Translation

A Pragmatic Concept of Translation*

Hsiu-Hwang Ho

*¥Z©ó¡m­õ¾Ç½×µû¡n²Ä¤@´Á¡A°ê¥ß»OÆW¤j¾Ç¡A»O¥_¡A1971¦~1¤ë¡C

1.The aim of this paper

Translation is a fairly common activity known to most of us. It is an activity of rewriting a body of discourse into another body of discourse upon the condition that between the original body of discourse and the resulting one, there exists some certain sort of relationship. What exactly this relationship is, I shall try to specify in this discussion.
However, I am not going to develop a complete theory of translation. What I wish to do is this: to consider translation as a linguistic activity which can be illuminated by the theory of meaning proposed and developed by Henry S. Leonard [1],[2],[3],[4], and thereon to outline a pragmatic concept of translation.
By a body of discourse I mean a string of signs related one with the other in such a way that the whole string taken as a unit can be used to signify some purpose or purposes intended by the person who utters the string.1
In this paper, I shall propose a definition, or rule, establishing the circumstances under which one may say that one body of discourse D' is a translation of another body of discourse D. A succession of different considerations will lead to a succession of refinements of that definition.

2.Interlinguistic and intralinguistic translation
Translation may proceed within a language, or it may proceed between one language and another language. As an example of the former, we may take the activity in which we rewrite a sentence into another sentence within the same language in order to facilitate understanding; as examples of the latter, we may take those activities of rewriting one of Shakespeare's poems into its story in prose, or rewriting Heine's "Du Bist wie eine Blume" into English. He shall call the former an intralinguistic translation , and the latter, interlinguistic ones.2
It is interlinguistic translation which primarily concerns us in this paper. Yet most of what we shall have to say will also apply to intralinguistic translation.

3.A Pragmatic account of translation
Pragmatics is defined as the science of the relations between signs and sign-users. It deals with problems such as the uses or functions of language, as, for example, when we classify the uses of language into cognitive, practical and esthetic ([3] or[4], p.16).However, it has a much wider scope than this. For instance, all the following affirmations fall within the scope of pragmatics:
(a) The word 'semiotic' is not used by Aristotle.

(b) In Denmark, people say 'Glaedelig Jul' when we would say 'Merry Christmas'.

(c) In Germany, people use 'Morgenstern'to denote a star.

(d) It is (pragmatically) undetermined whether the following is a sentence: 'Today is '.

To think about translation, when we rewrite a body of discourse D into another of body discourse D', we may, and often do, have in mind one of the following considerations:

(i)While D is in a certain language l, D' is in another language l1. For example, D is in English while D' is in German. That is to say, we want to translate a string of signs in English into another string in German.

(ii)While D is a certain literary style s, D' is in another style s1. For example, D is in poetic style, while D' is in prose style. That is to say, we want to translate a poem or a line of poetry into prose or into a sentence or sentences in prose form.

(iii)While D is difficult for some people x to understand, D' is easier for x to understand. That is to say, we want to translate a body of discourse which is difficult for a certain person or group of persons to understand into another body of discourse which is easier for him or them to understand.

But there are no pragmatic rules which require that something of this sort must be fulfilled in order that discourse D' may be called a translation of discourse D. For example, it is not always the case that we translate a string of signs more difficult for certain person(s) to understand into one easier for him (them). It could well be the reverse. We may on some occasions try to translate a paragraph easier for certain person(s) into one more difficult for him (them). Notably in war time's military correspondence, we try to send a code message instead of a message written in a plain language. In this case, we translate a paragraph (e.g., in English) which is more easily understandable for certain persons, namely, the enemies, into another for them more difficult one. Similar remarks hold for the other pragmatic considerations listed above.

There are, the several pragmatic considerations which are commonly practised but are not inviolable from a logical point of view. However, we can formulate a pragmatic rule which may not be violated in translation. As a preliminary formulation of this rule, we put down the following:

R1. A discourse D' is said to be a translation of another discourse D if and only if utterances of D' and of D would serve the same purpose or purposes.

In other words, D' is a translation of D if and only if they both exhibit the same use(s) or function(s) of language. For example, if an utterance of D would constitute a cognitive use of language, one of D' must also constitute a cognitive use; if one of D would constitute an esthetic use, one of D' must do likewise.

In connection with R1, I find Leonard's analysis of "purpose" into a concern and a topic of concern turns out to be especially illuminating. According to Leonard, an expression may be said to express a person's concern, it indicates his topic or concern, and we also say that it signifies his purpose. Expression, indication and signification are therefore said to be three different modes of meaning.3 In the light of this analysis, we may reformulate R1 as follows.

R1a. A discourse D' is a translation of another discourse D if and only if an utterance of D' and one of D would express the same concern(s) and indicate the same topic(s) of concern, i.e., if and only if they would signify the same purpose(s).

Two features of formulation R1a should be noticed. Each is a feature which will eventually necessitate some reformulation. (1) R1a is elliptical in two respects: it does not mention the languages, ,l1 and l, in which D' and D are framed. Neither does it refer the mentioned purposes to the presumed author of the original discourse D. The resulting discourse D' must be able to serve the same purpose(s) as the author of D intended to serve and to signify by his utterance of D.(2) R1a is so framed as to allow that both D' and D signify a multiplicity of purposes. The formulation suggests that all purposes that utterances of either one would signifies must be signified by utterances of the other. But surely this is too strong a demand.

Let us first attend to this second difficulty.

4.Essential purposes

In discussing purpose, Leonard makes some suggestive distinctions.4For example, he notes that a person may have many purposes in mind in the performance of a single action. Of these purposes, some may be said to be primary, others secondary; that is, some may be the main things that he want to accomplish, others, the minor things. Again, some purposes may be (relatively) immediate, others (relatively) remote; that is , some may be things that he wants to accomplish right away, others, after a while. Further, one purpose of an action may be said to be subservient to another purpose of the same action: purposed as a means to accomplish the other. Or the two may be said to be independent: that is, neither is purposed as a means to, or in order to accomplish, the other.

This analysis has a special significance to our pragmatic account of translation. For example, a body of discourse may signify more than one purpose of its author, but the signified purposes may not be of equal weight. In doing translation, the primary purposes(s) intended in the original discourse must be preserved as primary in the resulting discourse, while the secondary purpose(s) may or may not be changed or sacrificed according to the insight that a translator has when he considers in what way he can best signify in the translation the author's original primary purpose(s).

In short, a translator must distinguish between tolerably sacrificable purposes and essential (i.e., not tolerably sacrificable) purposes signified by a body of discourse D, and R1a must be modified so as to demand of the translation D' that it signify all of the essential purposes of D.

R1b. A discourse D' is a translation of another discourse D if and only if utterances of D' and of D would signify the same essential purposes.

The next few sections of this paper are concerned with some questions relative to the distinction between tolerably sacrificable and essentials purposes.

5.Subservient discourse

It seems of great illustrative to observe that one body of discourse may be subservient to another. Let us call the former a subservient body of discourse, or simply a subservient discourse; and call the latter a principal body of discourse, or simply a principle discourse.5 Among the various kinds of subservient discourse, there is a special kind that deserves our attention. I have in mind an exemplification. A body of discourse D1 exemplifies another body of discourse D if D1 designates a special case of what is said in D in order to help establish or clarify or facilitate the understanding of what is said in D. In this case, what primarily concerns the author of D1 is D rather than D1 itself.

For example, in ¡±15.1 of his Principles ([3]or [4]), Leonard begins the discussion with the following two paragraphs:

(A)¡qE How many figures appear in figure 5, below? It depends on what the word "figures" means. If "figures" means "shapes", then the correct answer is: two, a triangular shape and a rectangular shape. But if "figures" means "drawings", the correct answer is: three, the left-hand one, the middle one, and the right-hand one.¡r6

(B)¡qE The above question illustrates a kind of ambiguity that is liable to attach to all sorts of words used in dealing with language. Among these possibly ambiguous words are "letter", "word", "phrase", "expression", "sentence", and "sign". For example , one might ask, how many words appear in the first paragraph(A) of this section? The correct answer depends on what is meant by "word". If "words" means "dictionary words", then the correct answer is thirty-four. But if "words" means "printings of words", then the correct answer is fifty-four.¡r

The ending sentences of paragraph (B), beginning with 'For example,' may be said to be a subservient discourse with respect to the preceding sentences of that same paragraph. The former help to male the latter understood.

Suppose, now, we translate these two paragraphs (A) and (B) into another language, say, German. It could well be the case that the German translation (A') of the first paragraph contains not thirty-four dictionary words and fifty-four printings of words but rather twenty-eight dictionary words and forty-one printings of words. Now, in order to translate the second paragraph (B) into German (B'), we could not directly say that there are vier-und-dreissig dictionary words and vier-und-funfzig printings of words in the previous paragraph. We should say instead that there are acht-und-zwznzig dictionary words and ein-und-vierzig printings of words in that previous paragraph. In short, sub-discourse

(1)¡qE If "words" means "dictionary words", then the correct answer is thirty-four. But if "words" means "printings of words", then the correct answer is fifty-four.¡ris not synonynmous in any degree of strength with its translation

(2)¡qG If "words" means "dictionary words", then the correct answer is twenty-eight. But if "words" means "printings of words", then the correct answer is forty-one.¡r7

This example serves to point out that translation is not necessarily a synonymy-preserving mapping between two bodies of discourse, as is commonly and common-sensically assumed or understood to be the case. It is, however, not difficult to realize that sub-discourse (2) as occurring in discourse (B') with respect to (A') in German serves the same essential purpose(s) as sub-discourse (1) as occurring in discourse (B) with respect to (A) in English. In other words, they serve essentially the same illustrative purpose with respect to their previous paragraphs. Hence, we can say in conformity with R1b that translation is an essential purpose-preserving mapping between two bodies of discourse.

6.A counter-example to the pragmatic theory of meaning

Here we seem to face a problem. That is, we may throw a doubt upon the claim that pragmatic considerations can provide a sufficient condition for a theory of meaning. According to this theory, meaning (of a deliberate sign) is defined in terms of pragmatic characterization. A body of discourse has such a meaning (or, such and such meanings) if it serves for a certain author to accomplish thus and thus a purpose (or, thus and thus purposes). And two bodies of discourse have exactly the same meaning(s) if and only if they serve exactly the same purpose(s). But now we seem to encounter a counter example. While (1)in English and (2) in German serve essentially the same purpose(s), they, nevertheless, have different meanings.

In view of the problem resulting from the translation of (A) and (B) into (A') and (B') above, we may ask the question what makes the translation under consideration become not a synonymy-preserving mapping? It is obvious that (A) and (B) are on different language levels, and when we make a translation for (A) and (B) into (A') and (B') respectively, then (B') is not talking exactly about the same object as (B) is, although (B') is connected with (A') in exactly the same way as (B) is to (A). This observation may suggest people to think that our counter-example to the pragmatic theory of meaning is nothing but apparent, or rather it is only a very special case. For example, we can avoid the above difficulty if Leonard writes the following (C) instead of (B) above.

(C)¡qE¡KFor example, one might ask: How many words appear in the first paragraph of this section? The correct answer depends on what is meant by "words". If "words" means "dictionary words", then we get one answer; but if "words" means "printings of words", then we get another answer¡KI leave it to the readers to count out the exact number for each of the two answers.¡r

Then, the difficulty seems to disappear. But it would seem more appropriate to say that in this special case, the difficulty is suppressed rather than disappearing, because it always has the danger that the difficulty may in some situation or other reappear and come up onto the surface.

Because of this difficulty, people may tend to make a proposal to the effect that in a theory of translation let the language-level distinction be taken into consideration. They may even propose to say that if we have two bodies of discourse D1 and D2, D2 taking D1 as its subject-matter, then in our translation D1' and D'2, we shall not call D'2 a translation of D2, but rather that D'2 is an explication or illustration or something of this sort to D1' in view of D2. Take our previous example, (B') in German will not be said to be a translation of (B) in English, but an illustration of (A') in view of (B). But in practice this amounts to the fact that we translate (B) into (B') with the necessary corrections or amendments in the light of (A'). Therefore, it seems to me that we can call (B') a translation of (B) in the light of (A') as well.

In case we rewrite D2 into D'2 without a radical change of the content of D2, it seems without any harm at all to think that the one is a translation of the other. For instance, in (C) and (C') above, we make no change as to the content of (C) in (C'), (C') can be said to be a translation of C, despite the fact that (C') is talking about (A') rather than about (A) , which (C) is talking about. However, I am also inclined to think of (B') in our example as a translation of (B) although in (B') we did make some substantial changes in the content of (B), namely, we change "thirty-four" into "twenty-eight", and "fifty-four" into "forty-one".

7.Other examples of non-synonymous translation

There are cases of translation in which we make a radical change in resulting body of discourse with respect to the content of the original body of discourse. For example, if Quine lectures at Harvard University to make a distinction between the use and mention of a linguistic expression, he may say something as follows:

(3)¡qE¡Kas we know very well that Boston is a big city, but 'Boston' is a six-lettered word.¡r
Now, suppose a Japanese translator wants to translate (3) into the Japanese language, he might render it in this way:

(4)¡qJ¡Kas we know very well that Tokyo is a big city, but 'Tokyo' is a two-word phrase.¡r8 He does this because in the Japanese language there is no way to talk about the alphabetical letters comparable to that in English.

There might be people who would like to think that (4) is not a translation of (3), but, rather, that if (3) is subservient to, say, D, (4) to D'9 , then (4) is an illustration of D. The main reason for refusing to call (4) a translation of (3) is that (4) and (3) talk about entirely different terms, one about the city Boston and the word 'Boston', and the other about the city Tokyo and the phrase¡qJ 'Tokyo'¡r. But since (4) serves the same essential purposes as (3), R1b would justify our calling (4) a translation of (3).

8.Hypothetical intention

We have followed Leonard in saying that to ask what an utterance means as a deliberate sign is to ask what its speaker was intending to accomplish by making that utterance. Now, in talking about translation we need to make a minor change in the above affirmation. Let us say that to ask what purpose(s) a body of discourse, as a string of deliberate signs, serves to single out is to ask what its author would be intending if he should make that body of discourse.

To take the example of Quine's lecturing in Japan, what we claim is the following. If he should go to give a lecture in the Japanese language at Tokyo University, in making a distinction between the use and the mention of an expression, it is very likely that he would take Tokyo (and¡qJ 'Tokyo'¡r) rather than Boston (and 'Boston') as an example. Hence when we translate (3) in English to (4) in Japanese, although these two strings of signs are not synonymous, they can, nevertheless, serve essentially the same purpose(s).It is true that there may be more than one, or even many, bodies of discourse other than (4) in Japanese which can be used to serve the same purpose(s). But our point is to think of the translator as being able in one way or another to look into the mind of the original author., and discover what he would be most likely to say or to write under that context in question. If Quine should lecture in the Japanese language at Tokyo University, he would probably take Tokyo as an example; if he should lecture in the Chinese language at National Taiwan University, he would probably take Formosa or Taipei as an example. Although we do not possess any way by which we can effectively say exactly what body of discourse an author would be likely to use of he should write or speak in a certain circumstance, nevertheless, we must choose a string of signs which can do essentially the same job as the author's original string of signs. That is to say, these two bodies of discourse must be able to serve essentially the same purpose(s) as we defined above. If we say that a body of discourse D uttered by an author serves to signify the purpose(s) intended by him, then let us say that a translation D' of D signifies the purpose(s) hypothetically intended by the original author.

Now, we may write our pragmatic rule for translation as the following:

R1c. A body of discourse D' is a translation of another body of discourse D if and only if D' signifies the purpose(s) hypothetically intended by the author of D while D signifies the purpose(s) intended by him.

9.Linguistically dependent elements and translatability

We have so far exhibited that in the case of subservient discourse there may be purpose-preserving but not synonymy-preserving translation. And furthermore, we have extended the notion of purpose to cover those case in which the so-called hypothetical intention is involved. However, these are not the only cases in which translation preserves purpose but does not preserve synonymy. The reason that we take those subservient discourses into consideration is that in that circumstance we are able to determine more easily what purpose may be intended or hypothetically intended by an author in uttering a string of signs. We do this simply by referring to the discourse to which this one is subservient. But how about a string of signs which is not apparently subservient to any other string of signs in a particular context? The situation, I think, is similar.

In order to see the purpose-preserving but not synonymy-preserving translation more convincingly, let us take an example from the emotive use of language. Suppose in a certain nlaguag l, the sentence

(5)¡ql He is a pig.¡r
has a very bad derogatory implication, but the sentence

(6)¡ql He is a dog.¡r
has little or even no emotive association at all. Suppose, further, that in another language l1, the reverse is the case, that is

(6')¡ql1 He is a dog.¡r
has derogatory implication, while

(5')¡ql1 He is a pig.¡r
has little or none at all, Now in order to translate (5) in l into another sentence in l1, I would think that we translate it into (6') rather than into (5'). Because if we translate (5) into (5'), then, it may read somehow like

(7) He is maximally satisfiable class of wffs. which makes little sense in that context in l1.

As we may have suggested above, the translation of any body of discourse which talks about something that is at-least-in-part dependent on the linguistic element creates the difficulty we are talking in this paper. But this does not necessarily create an unconquerable task as we showed above. In some cases, however, we may not be quite sure exactly how much we can do. For instance, ambiguity is something which is associated with linguistic elements. We may have an ambiguous expression in this language but no equivalent expression in another language. In case we want to make use of the very fact of ambiguity in a string of signs N, we can hardly find a legitimate translation N', in another language in which no string of signs has the corresponding ambiguity. The following is an example. A British or an American student who is puzzled by philosophy may utter something like this:

(S)¡qE What is mind? no matter; what is matter? never mind.¡r
This string if signs says something which conjoins in one way or another the following two strings:

(S1) No matter what mind is; and never mind what matter is. And

(S2) What is mind is not matter, and what is matter is not mind.
But neither (S1) nor (S2) taken separately signifies what (S) signifies. The interesting point here lies in the fact that in (S) the author artfully uses the ambiguity associated with words like 'no matter' and 'never mind' not occurring isolatedly but within that context. If in trying to translate (S) into another language, say, Chinese, we may not have the same ambiguity associated with the corresponding words. Or to put it in another way, we may translate (S1) into a certain (S'1) and (S2) into a certain (S'2) without any difficulty, but (S'1) and (S'2) cannot be conjoined into an (S') in the same way in which (S) had been done. Therefore, what can be signified by (S) may not be signified by (S'). That is to say, (S) and (S') may not signify the same purpose. Hence (S') is not a translation of (S).

But how about the following possibility? Suppose that instead of 'no matter' and 'never mind' in Chinese, we have 'no Y' and 'never Z' which mean "no matter" and "never mind" respectively. But when we take 'Y'and 'Z' isolatedly, they mean "analytic" and "synthetic" respectively. Consider the following sentence:

(S")¡qc What is z, no Y; what is Y, never Z.¡r
I would say that this could signify the purpose (s) hypothetically intended by the author of (S). That is to say, if the author is to utter in the Chinese what is said in (S), he will say (S") rather than something else. Hence, according to our R1c, (S") is a legitimate translation of (S). This is not synonymy-preserving, but it is purpose-preserving.

At this point, we may raise the question: Is there, then, anything which is not translatable? The answer seems to me in the positive. Actually, the example we just gave is entirely fictitious. In the Chinese language there is no way to render (S), even approximately, as we suggested above.

Another example is this. If we want to make use of the ambiguity in¡up v q v r¡vwritten in so-called standard notation, there is no way to render it into Polish notation and still preserve the ambiguity therein.

10.Culturo-historical aspect of a state of affairs

We may have, so far, created an impression that the problem of translation we are dealing with comes from the fact that different language possess different peculiarities. But it is rather difficult to say to what extent the relevant peculiarities are exclusively linguistic. For example, we might be inclined to think that we may translate a string of signs in the language used by this group into another string of signs in another language used by another group. But this statement is not clear in many respects.

For example, in the island of Tahiti young people of the opposite sex express their affection not by touching their lips, buy by touching their noses. Let us call it nunu. Now, suppose a line of a poem written by a Tahitian poet t run somewhat as follows.

(T)¡qT Romeo nunus Juliet gently without a word of sorrow.¡r
when we try to translate (T) into English, we may have a misgiving as wheher to render it into

(T')¡qE Romeo kisses Juliet gently without a word of sorrow.¡r
or not. We cannot translate (T) into, e.g.,

(T")¡qE Romeo presses his nose on Juliet's gently without a word of sorrow.¡r
because the phrase 'presses his nose on' in English does not carry the impression of expressing his affection. Of course, the simplest and commonest way to do this is to transplant the word "nunu", perhaps written in italic, into the vocabulary of English. But this is not a translation between the Tahitian language and English. It involves the enlargement of English vocabulary by transplanting the word 'nunu' before we carry out the translation.

In this case, the difficulty seems to lie in the fact that between these two groups of people, namely, the Tahitians and, say, the Americans there is no common state of affairs at that point. Although pressing of lips and pressing of noses taken physically are, or could be, state of affairs common to them; but when taken otherwise, usually culturo-historically, this is not the case.

This example, I hope, may help indicate the following claim. That we should not think of the so-called states of affairs as absolutely independent of the culturo-historical background of a certain group of people. States of affairs are construed via the employment of language. This makes a state of affairs being talkable or enumerable and what not. But the very act of construing a state of affairs via the use of language casts a cultural veil over it. Therefore, I would think that a state of affairs is not purely physical, it is also in a sense cultural.

However, this is not to be taken as implying that we cannot isolate the purely physical aspect of a state of affairs, e.g. pressing noses, from the culturo-historical aspect of it, e.g., expressing affection by pressing noses. What we claim here is that in calling a certain state of affairs, say pressing noses, by a certain name, say, nunu, we already subsume the state of affairs under our cultural casting. And different cultural groups may construe a state of affairs in quite different ways.

11.Unit of translation and equi-legitimate translations

In doing translation, we proceed by rewriting a body of discourse into another body of discourse. Now a question may be raised as to how large a body of discourse we must take as a unit to do translation. Do we translate word by word, or do we translate sentence by sentence, or even paragraph by paragraph, and so on? There are people, notably Mates ([5],p.112), who affirm that the unit of translation is arbitrary to the extent that the translator thinks which way will produce the most satisfactory result. Hence, the unit of translation may be a word, it may be a sentence, or, it may be a paragraph, and so on. Although this affirmation does not conflict with the position we take here, however, we can have a more definite answer. Since we have a pragmatic concept of translation, or more specifically, since we affirm that translation is a purpose-reserving mapping between a body of discourse and another body of discourse, and since the purpose intended by the author can only be conveyed successfully through the use of a sentence or a sentence equivalent, we can conclude that it is a sentence or a sentence equivalent that is the simplest unit of translation. However, as can be inferred from what we have said so far, we frequently take into account , or make reference to, other more extensive units, such as a sub-paragraph consisting of several sentences or a whole paragraph, or, sometimes, even several paragraphs. This is especially the case when we try to determine the purpose or purposes hypothetically intended by an author.

When we talk about translating a body of discourse D into another body of discourse D', we call D' a translation of D rather than the translation of D. We do this because it often happens that between two different languages, for a certain body of discourse in one language, there may be more than one body of discourse in the other language purpose-preservingly corresponding to it. Generally speaking, to a certain body of discourse D in l, there may be D'1, D'2, D'3,...,D'n (n¡Ù1) in l1 corresponding to it as legitimate translations. Let£Fbe a set consisting of all the D'1 to D'n, then we may call every member D'I(i¡Øn) of£F an equi-legitimate translation to another member D'j (j¡Øn and j¡Úi) of£F.Or in other words, we may simply call any member D'I of£Fone of the equi-legitimate translations of D.

However, in saying that D'i and D'j are two equi-legitimate translations of D according to the above definition, we are not saying that D'i and D'j are equivalent in the sense that they preserve exactly the same purposes intended in D in exactly the same manner. That is, every single purpose intended in D which is preserved by D'i is at the same time preserved by D' j in the same way, and vice versa. It may well be the case that there are some (non-essential) purpose(s) preserved by D'i but not preserved by D'j , or some purpose(s) preserved by D'j but not by D'i.But we do claim that they must preserve essentially the same purpose(s), i.e., those unsacrificable purpose(s), which is (are) intended or hypothetically intended by the author of D, in order to be the members of£F. In short, each member D'i of£Fshould thus preserve the purpose(s) of D in order to be a translation of D at all.

12.Full and partial translation and grades of translation

Along with the line of reasoning given in this paper, we can distinguish between a full translation and a partial translation of a body of discourse in the following way. Let D be the original body of discourse, p1, p2, p3,...pk (k¡Ù1) be the unsacrificable purposes intended by its author in D, and D' be a translation of D. Suppose that D' preserves every item of p1 to pk, then D'is a full translation of D; otherwise, if D preserves at least one but not all of p1 to pk, then it is a partial translation of D. Hence we can establish the notion of the degrees or the grades of translation.

This analysis is especially helpful when we deal with a body f discourse which is intentionally ambiguous, i.e., with more than one unsacrificable purpose intended in it.

But the distinction between a full and a partial translation is always relative according as what a series of purposes the author of a body of discourse will count as essential or unsacrificable. For example, in section 9 of this paper, (S")can be a full or a partial translation of (S) depending upon how many purposes the author of (S) will count as essential. But neither (S1) nor (S2) will be a partial translation, let alone a full translation, of (S) as it is taken conventionally.

References

1.Leonard, Henry S. "Authorship and Purpose"Philosophy of Science, vol.26, no.4 (1959), pp.277-294.
2.______ "Interrogatives, Imperatives, Truth, Falsity and Lies."Philosophy of Science. vol.26, no.3 (1959). Pp.172-186.
3.______ Principles of Reasoning. Dover Publications, Inc. New York, 1967. Revised edition of [4].
4.______Principles of Right Reason. Henry Holt and Co., New York, 1957.
5.______ "Synonymy and Systematic Definition." The Monist. vol, 51, no.1 (1967)
6.Mates, Benson. "Synonymity." In Linsky, L., Semantics and the Philosophy of Language. The University of Illinois Press, Urbana, 1952, pp111-136.

Hsiu-Hwang Ho

Stanislaus State College


*The author wishes to dedicate this article to the late Professor Henry S. Leonard, and thereby to express his indebtness to him. Professor Leonard read two earlier versions of this paper and expressed some invaluable criticism. However, not every point he discussed is fully accommodated in the present version. Hence, the author alone is responsible for the inadequacies or mistakes that may be found in it.
The paper was written in 1968 and then was read at the philosophical club of Michigan State University.

1.Cf.[3] or [4], ¡±¡±14.3-14.6

2.For a definition of a language, see [5].

3.Cf.[3] or [4],. ¡±¡±14.3-14.6; also [2], pp.175-181.

4.Cf. [3] or [4], unit12, [2],pp.175-177, and [1], pp.278-283.

5.This distinction is made in a relative sense.

6.I use '¡ql...¡r'to indicate that '...'is written in language l. And I use 'E', 'G', 'C', 'J', and 'T' to stand for English, German, Chinese, Japanese, and Tahitian, respectively.

7.On strength of synonymy, see[5], ¡±9.

8.In Japan, people write the name of Tokyo as 'ªF¨Ê' which consists of two words.

9.We do not specify D here, but it may be thought of as a paragraph talking about the use-mention distinction. And D' here being a translation of D.