5: "On the whole our modern cultural preference is for a fairly free translation. A word for word translation is favoured only for very limited purposes, for example for the use of beginners in a language, who may need a key which will enable them to follow the exact wording of the original...a version of any book is not necessarily consistent throughout: while generally literal, it may at certain points quite suddenly take a leap into what appears to be a very free rendering."
6: "For - and this is my principle argument - there are different ways of being literal and of being free, so that a translation can be literal and free at the same time but in different modes or on different levels...The simple comceptual distinction between literal and free is not sufficiently flexible to formulate the more complicated set of relations which obtain in many actual texts."
7: "...as this study may show, truly 'free' translation, in the sense in which this might be understood by the modern literary public, scarcely existed in the world of the LXX, or indeed of much of ancient biblical translation in general...A sophisticated study of the LXX, at least in many books, has to concern itself much of the time with variations within a basically literal approach: different kinds of literality, diverse levels of literal connection, and various kinds of departure from the literal."
9: "It is important therefore that we disengage the identification of the methods of literal translation from judgements about other aspects of the style and methods of those who have practised it."
11: "...the question of literalism in translation is seldom or never absolutely provable by the mere juxtaposing of two texts, say one in Hebrew and a rendering in Greek. The decision and the assessment of the evidence is not settled by the mere existence of the two texts: it depends on a semantic judgement, the setting forth of a semantic path which may reasonably be taken to have led from one text to the other."
12: "...wherever there is a Hebrew text that (subject to a proper understanding of the translation technique followed) show substantial semantic agreement with the Greek rendering, the probability is that the Greek has resulted from a fairly literal translation of that text."
13: "Clearly, a literal rendering must show a fairly close verbal correspondance with a Vorlage extant or reconstructable. Equally, however, a rendering, to be usefully classified as a 'free' rendering, must have some recognizable semantic path that leads from the Vorlage to the rendering...Unless there is some formal/semantic link between Vorlage and rendering, the versional phrase cannot usefully count as either literal or free translation: it cannot count as translation at all."
14: "If a rendering is both 'free' and also seriously mistaken, there is not much we can do with it: unless there is some literal element in it, some element linking it with the formal character of the original, or some general semantic rightness about it, we cannot usefully classify it, it is not in essence different from a wild guess."
15: "It seems therefore that we end up with three types:"
(a) "free" renderings which state more or less correctly the general purport of the original text
(b) literal renderings which also give an adequate semantic rendering of the original
(c) literal renderings which, while their semantic indication is far from being an adequate indication of the meaning of the original, nevertheless show a close and understandable relation to the form of the original.
15: "A modern man of letters, translating (let us say) Dante for the modern public, may opt for a free technique of translating, but he would be offended if it was supposed that he did this because he did not know the meanings of the actual words of Dante's Italian. In this sense free translation is a literary embellishment, perhaps an essential one, but one superimposed upon a previously existing certainty of basic underlying linguistic understanding. In the ancient biblical trarnslation it is often the opposite: one of the main factors that determine the bahaviour of translators is the sheer obscurity of certain phrases and their own lack of certainty about the meaning of elements in the text."
15-16: "Thus the fact that books like Job and Proverbs have often been noted for the 'free' style of their Greek version can rightly be connected with the fact that these books are near to the edge of the biblical canon and less central to the structure of religious doctrine. But, if this is true on one side, it is equally proper to note that in these books thet Hebrew diction itself was often very obscure, and that some fair proportion of the freest renderings seems to coincide with very obscure phrases of the original."
Come back to this. He hasn't made his point well enough to lean on this.
16: "But this same starting-point, a deep obscurity in the original, may serve as generating force for both of the two supposedly contrary tendencies in translation. If a text is really difficult and obscure to the translator, he may opt for free translation, making a general estimate of the total meaning, or simply guessing at it, and ignoring the details; but he may also do the opposite, and decide to give a precise impression in Greek of the detailed form of the Hebrew, leaving it to his readers to work out, if they can, what the general purport of this may be. What is more, the translator may combine the two, providing a rendering based on a precise relation to the Hebrew at certain points and therefore in these aspects literal, and filling in the rest with guesswork or very general interpretation."
Excellent
16: "Once literal techniques are worked out, they tend to spread and find their way into every kind of locution...It is commonly said that every translation is also an interpretation. In the context of ancient biblical translation, this remark is a highly misleading truism...Or, perhaps, we may better say that the process of translation may involve two quite different sorts of interpretation, which are so different as hardly to deserve to be called by the same name."
17: (see above note) "The first is is a sort of basic syntactic/semantic comprehension of the meaning of the text: if you have a piece of writing in Hebrew which you cannot understand at all you cannot translate it...The other type of interpretation lies on a higher level: it begins only after these linguistic elements have been identified...The translator, while he is bound to carry out the first kind of semantic interpretation and cannot act at all without it, is free to choose how far he enters into questions concerning the second kind of interpretation."
18: "A translator is bound in some degree to be an 'interpreter' on this level; but he is not at all bound by the nature of his task to be an interpreter in the sense of an elucidator of the content...The point of this for our purpose is as follows: far from it being the case that every translation is also necessarily an interpretation, there could be points in some ancient translations of the Bible where one of the main motives was, if we put it paradoxically, to avoid interpreting."
20: "The following appear to be distinguishable modes of difference between a more literal and a less literal rendering of a Hebrew text:"
1. The division into elements or segments, and the sequence in which these elements are represented.
2. The quantitative addition or subtraction of elements.
3. Consistency or non-consistency in the rendering, i.e. the degree to which a particular versional term is used for all (or most) cases of a particular term of the original.
4. Accuracy and level of semantic information, especially in cases of metaphor and idiom.
5. Coded "etymological" indication of formal/semantic relationships obtaining in the vocabulary of the original language.
6. Level of text and level of analysis.
22: "Generally speaking, it is not possible in any text, in any language, to make even basic identifications of words without some attention to their context, which is the sole resource availible to select between the multiple possible values of the signs."
26: "The division of the text into elements or segments, commonly at the word level rather than the phrase level, is then a normal feature of the more literal style in translating."
29: "A translation may amplify the text it is translating, adding to it considerable amounts of new material."
29-30: "Quantitative divergence from the original in either direction means loss of literality. A literal translation will express only the linguistic elements that are present in the original, and will express all of them. The measure for this is of course semantic: there is no way in which a Greek text can be merely quantitatively equivalent to a Hebrew text, except that it expresses meanings that stand for the meanings of each element in Hebrew."
30: "When large additions are made, these can sometimes be regarded as free composition. The version is then a mixture of two things: it is in part a translation of the original, in part free writing, meditation, commentary, midrash and the like. When this is so we may find that the version combines the literal and the free: where it is actually translating the original text it may be quite literal, but the additional material is free."
30: "And certainly if the material is pure free compostion, having no base in the original text, we hardly need to concern ourselves with it...The same is true for the occasional added phrases or sentences of the Greek Proverbs...Thus expansions are often not mere additions, they are exegetical provisions of context."
31: "Firstly, the exapansionism of the Targums was made tolerable because the Hebrew text itself was also there in the same community; and even if not all could understand it, or understand it well, it remained present as a measuring-standard and as the acknowledge source of ultimate authority...In the Greek world the opposite situtation prevailed. The amount of amplificatory matter added to the Greek text, whether by Jews or (later) by Christians, was very slight."