Errors in Translation: A Tool for Linguistic and Socio-cultural Competence

One of the important insights in recent translation studies research is that errors are a consequence of both linguistic and cultural misconceptions. In recent years translation studies have become increasingly involved in a quest for explanations of all phenomena associated with meaning interaction and a lot of detailed research has been attempted at most translation stages. Whole solid books have been written on specific topics but whole solid books can not include all the variations that different manifestations of language might adopt. As far as we understand, these theoretical ideas have not normally been applied to translations so far and when they have, explanations and descriptive interpretations given do sound rather artificial and unsatisfactory. We intend to propose an analytic approach to solve problems on translation based upon the principles of identity or equivalence, the main ideas of which might be suitable both for research and tuition purposes. In this paper, clear and cut distinctions between canonical and non-canonical expressions, collocations and idiomatic expressions are summed up succinctly both for language explanations and translation analyses, due to the fact that a good number of realizations belonging to the Gray Areas (GA) of language may arise from recurrent combinations of specific types of combined lexical items. The resultant lack of interaction between L1 propositions and L2 representations, identical or not, is often explainable and clarified by the Error Analysis (EA) method. Data collected and analysed here have been chosen at random.


Introduction
Numerous debates have been produced as far as translation proposals are concerned along the history.It has often been said that no text can be analysed by itself, far from the culture in which it has been conceived, far from the period in which it has been written.Thus, contemporary scholars debate on the fact that at least these three factors are to be taken into account: linguistic, cultural and time dimensions.However, of the three the first two were given more attention from a theoretical point of view and this gave rise to two different approaches upheld in the first case by Nida, Wilss and Steiner among others and the second by A. Lefevere, Even Zohar, Vermeer, Nord, Höning, etc.The first group, following the 'universals' linguistic theory insist on the fact that there exist a set of universal syntagmatic structures applicable to any language indistinctly.The second seeks support to resolve formal linguistic problems due to the fact that there exist intercultural factors which are far different from one culture to others i .This dichotomy, however, as usual as far as human production is concerned, though perhaps more elaborated, is nothing new as for devoted scholars to this field of research such as St. Jerome, Luther, Dryden, Montaigne, Tytler, Schleiermacher, Beatriz de Luna, Isotta Nogarola or illustrated persons who have also dedicated a good part of their lives to the art of translation such as Goethe, Mallarmé, Cortázar, Menéndez y Pelayo, Gómez de la Serna, among others, have had this double conception in their minds as the most relevant concerns within this field of studies (Oro: 2001:1-12).
During the XIIth and XIIIth centuries it was the cosmopolitan Toledo's Translators School (a set of studiers who have translated scientific and philosophical Arabic works ii to Latin and Spanish) a prestigious scientific center grouping translators from the three relevant religions of the period: Islamic, Hebrew and Christian.The work done there at that time embraces science, literature iii but also religious works; for example, one important achievement was the translation of Mi'ray or Mahoma's Scale iv both to Castilian and Latin, or the Coran's translation by Herman the Dalmat, the English Robert of Chester and Spanish Pedro de Toledo.It is compulsory to mention Abū l-Walīd Muhammad ibn Ahmad ibn Muhammad ibn Rushd (Arabic: en ‫ﻮ‬ ‫أﺑ‬ ‫ﺪ‬ ‫اﻟﻮﻟﯿ‬ ‫ﻣﺤﻤﺪ‬ ‫ﻦ‬ ‫ﺑ‬ ‫أﺣﻤﺪ‬ ‫ﻦ‬ ‫ﺑ‬ ‫ﻣﺤﻤﺪ‬ ‫ﻦ‬ ‫ﺑ‬ ‫ﺪ‬ ‫,)رﺷ‬ known as Averroes v (1120-December-10,198), the best studier of Aristotle at that period, but there were also other important Arabic scholars such as the Persian, Avicenna vi , al-FärÄbï vii or the Jewish, philosopher and physician, Moses Maimonides.This important work had a double dimension: the translation of works of the Islamic culture proper and the recuperation of classical works lost in Occident which has been recuperated through Arabic translations.
One of the main contributors to the double dichotomy and other translation principles through which contemporary scholars try to set their approaches to this hard discipline of studies was Saint Jerome (4 th century).His methodology has set the fundamental laws, especially that concerning sense.However, even though for him the act of translation especially for general texts would not consist only in transposing words from one language to another (i.e. this process needs to go further than that, translation becomes an act of interpretation -'ut interpres, ut orator'), he would not apply this principle to Sacred texts (for example, the Sacred Writings, in this case as they perform an act of faith the possibility of interpreting is not permitted; thus, he claims for identity.Perhaps the weakest point of his doctrine was to distinguish between two types of texts for translation analysis: sacred and profane, and attributes different principles to their design.
No doubt, this was an important contribution as for he proposes that there is a need for an adequation of the translation process to text types.His influence in posterior scholars and translators is a fact, for example the principle of literalness or literalism for sacred texts was the basis approach that Fray Luis de León in his translation of the 'Cantar de los Cantares' or Martin Luther's Bible version has followed.
It was not until the XVth.century when another interesting proposal for this field arrives.We are referring to the principles set by Juan Luis Vives viii , which in spite of being left soon aside, were retaken by German Romanticists during the XIX century.Vives makes a clear distinction between translation theoretical approaches and the practical situation when translating; besides, he realises that different languages are at the same time very similar and very different and he feels that even though translation is a maxim fixed and tacitly accepted, he understands that languages with peculiar and specific characters do present certain barriers impossible to pass through for translation interaction.He is referring to the canonical and non-canonical language representations proper.He concurs with San Jerome and Cicero on the fact that the translation process is a thoughtful and meditative activity in which the domain of liberty moves between the language to which meaning is transferred and the sense of the text which one pretends to translate.Luis Vives (1492-1540) is, no doubt, aware of some limitations which often might appear in this process: lack of expressions between the languages in question and a possible deficiency to understand the original sense proposed.He is also aware of the fact that proposals based on ideal norms for translation purposes do not exist in spite of the existence of certain general norms and conditions attributed to the field, such as: knowledge of languages and technical language, knowledge of the topic, culture, etc.

Sampling of Errors
The central core of this analysis will be essentially based in the comparison of certain units from 1961 Penguin edition of Brave new world and 1969 Hernández's translation into Spanish.Posterior editions to the above mentioned, (for example, 1980 BARCELONA, PLAZA & JANES) will be referred to from time to time due to the fact that it includes various changes, noticeable changes, from other editions.
As far as Hernández's translation is concerned, it is interesting noticing that there are variations from one edition to another and that there are variations comparing with other editions as well As can be observed in the figs.below differences between L1 and L2 have a double appearance: formal (i.e. they allow to recognise linguistic alteration and the typical and atypical world referents' representations) and as a result semantic (i.e.They affect meaning as being transferred to L2).
Formal differences have often been categorised into 3 main types of errors: (O) omissions (omitting a unit of meaning); (A) additions (adding one unit of meaning) and (S) substitutions (substituting one unit of meaning).This often involves a lack of linguistic interaction which provokes a certain prochain or remote distance between the two languages and the world referents or mental representations (i.e concrete or abstract references) often referred to as grammatical (GR).This misuse of grammatical devices between the two languages is often the result of: (M) morphological (changing word-classes); (L)Lexical variation (word, phrases, verbal tense confusion, etc) from conceptual of basic meaning to denotative meanings, or even from hyponym to hyponym; from super-ordinate to hyponymy, or on the other way round; SY) syntactic alterations from one language to another (changing unnecessarily any piece of syntax; for example, unnecessary word order change); (P) punctuation; (G) Graphological (Printed errors) and (SE) semantic (using units semantically related instead of the proper ones: for example a superordinate term instead of a hyponym).
All these aspects as a result would lead the interaction between L1 and L2 representations to grammatical and semantic distortion.
In general, the above categories are referred to by various scholars, following House (1977) and her model intent for translation quality assessment, as substitutions, which together with omissions and substitutions would constitute one of the three subtypes of the second part of her dichotomy errors classification (covert ix and overt) x usually found in translations according to her and her followers from a formal point of view at first sight.
To make a succinct comparison with various editions: the original writes cardinal points beginning with a small letter, but the Mexican DIANA, the Spanish Plaza & Janes and Collección Millenium (published by the Journal 'El mundo' use capital letters to translate them and the Galician Edicións Xerais uses small letters like the original.The same happens with the word 'Sauvage [salvaje]' which appears with a Capital all along the original version.The following tables exemplify some errors at the beginning of the novel which might alter the message given in L1.Variations also exist between or among different editions; some are clear misspellings, others change from one edition to other, some for worse and some for better and yet others remain unalterable edition after edition.

A Linguistic Approach. Canonical versus Non-Canonical Expressions
Both translation and linguistics are condemned to be integrated, as for translation cannot survive on sociocultural justifications.
On the one hand, linguistics is the study of our knowledge of language: what it is, and how we acquire and use it; the study is pursued through the construction of grammars; that is hypothesis about this knowledge and how we come by that knowledge and use it to think or communicate.The knowledge of language is not monolithic.It is usually divided into our knowledge of vocabulary and our knowledge of how we combine that vocabulary into sentences: id est. the lexicon and the rules of formation.Some of this knowledge is easily accessible in most languages.Translation, on the other hand, must be understood as a linguistic mechanism to process meaning identity from one language to another, through the study of their linguistic behaviour as a whole but out of the sum of the different parts that compose the pieces of language to be translated.It is our intention in order to integrate both fields of research to combine information from several different sources distinguishing between canonical patterns and non-canonical patterns of lexical domain.The former tend to be universal and the latter tend to be more exclusive.To make a proposal concerning translation we will include brief references to different expressions embraced under the super-ordinate term 'grammatical expressions', also referred to as 'the gray areas of language', idiomatic expressions, set phrases, etc..This creates a fabric of ideas which range from formal canonical appearance in its diverse varieties to semantic uniqueness far from the meaning of the parts that very often compose these pieces of language object of linguistic analytic and descriptive interest.It is also our aim to combine linguistic behaviour with the translation process in order to see if adequacy is possible and if diversity should be avoided.Even though the different states are still in a situation of development, we can conclude that only some of these expressions are to be considered lexical items proper.The majority of them follow normal formal canonical patterns and meaning may rouse raging from the transparent or semitransparent meaning of their constituents to the totally opaque, or they keep on going with their conceptual meaning as transparent realisations, i.e. a maxim which seems to be constant in great many linguistic processes (cf., for example, put at its simplest, verbs acting as full verbs or as auxiliaries).However a good number of expressions behave as lexical items proper and thus they are to be treated independently for language internal understanding but especially for translation comparative development, especially opaque realisations.Opaque realisations cannot be deduced adding together the meaning of the parts, i.e., they cannot be deduced linguistically, as the meaning goes beyond its conceptual meaning or through other extra-linguistic patterns, which are to be delimited.The collocational process is a general linguistic process in Language behaviour that can be split into: The associational process which ranges from loose to tight and derives into proper collocations or idiomatic phrases, normally set phrases, and a previous step for compounding.Then, it is important to have in mind that the main problem a translator has to face, once he/she has decided the stuff he is going to deal with and the appropriate materials and methodology to be employed are the linguistic differences at different levels.In general, most of them should not present any problems but some do, for example expressions considered to be unique or expressions the semantic extension of which has evolved differently in both languages due, for example, to socio-cultural inferences or to technological advances.That is a real fact, and of course we are not thinking of the awareness of expressions based upon canonical patterns, well, by canonical patterns we mean, for example, -as far as simple sentences are concerned: those following the structure: subject, predicate and International Journal of Applied Linguistics & English Literature complement, -as far as lexis is concerned those lexical units under their conceptual meaning, not presented under other types of meaning that are the result of semantic extension and/or linguistic arrangement.
In sum, taking this into account, for understanding translation problems it is useful to consider two types of expressions which are to be treated independently when translating.
The first, canonical expressions xi should not present problems for the translation process neither with lexical items nor with grammatical ones as it is illustrated in fig. 1 below, which corresponds to sentence pattern IV.
Opaque realisations cannot be deduced adding together the meaning of the parts, i.e., they cannot be deduced linguistically, as the meaning goes beyond its conceptual meaning or through other extra-linguistic patterns, which are to be delimited.

Canonical and Non-Canonical Expressions
The collocational process is a general linguistic process in language behaviour that can be split into: The associational process which ranges from loose to tight lexical units and derives into proper collocations or idiomatic phrases, normally set phrases or compounding.In most occasions they constitute the first step for the compounding resultant state (well known -àwell-known à wellknown; hard workingà hard-working à hardworking; book caseà book case à book-case, etc. mother in lawà mother-in-law; bull's eye), which is the result of combining two or more words to form a single unit.Then, it is important to have in mind that one of the main problems a translator has to face, once he/she has decided the stuff he is going to deal with and the appropriate lexicon and rules of formation to be employed, is the typical and atypical deviations from one language to another.In general, the use or selection of canonical expressions should not present any problem if the translators should have enough linguistic background and control of their native languages.
That is a real fact, and of course we are not thinking of the awareness of expressions based upon canonical patterns.By canonical patterns we mean, for example, one the one hand, those following the most common structural clausal or sentence patterns; i.e. subject, predicate and complement; or the normal phrase structure, i.e. head and modifier(s); in general, processes which are recognised in most languages, which tend to be universal, at least from a semantic point of view and which are not far from their formal representations.On the other hand, those referring to those lexical units xii under their conceptual meaning in their surface structure representation or other clear and cut types of meanings which tend to be universal to realise linguistically most world concrete or abstract objects or ideas, which are not considered to be unique or exclusive of one culture or of a particular community.We can probably exclude those lexical units which have acquired a very specific meaning by semantic extension and/or linguistic arrangement, full or idiomacities and irregulatiries.
In sum, taking this into account, one can deduce that there are two types of expressions which are to be treated independently for linguistic and translation theoretical and practical analysis.The first, canonical expressions, i.e. those expressions -from a lower to a higher rank-which follow general formation, realisation and functional patterns, either alone or in combination, should not present problems for the translation process neither with lexical items nor with grammatical ones as it is illustrated in fig. 1 below, which corresponds to sentence pattern IV, due to their straightforward reference to world objects, concrete ideas or abstract ideas world-wide developed.They normally range from basic and transparent communicative messages to semi-transparent interpretable realisations.They constitute essential basic communicative constructions for all or most communities.The second, non-canonical expressions are far more complex, they are affected formally, morphologically and syntactically, but essentially semantically.Put at its simplest, this specific behaviour of language can be viewed as a special type of indirect speech, and it constitutes a violation of the behaviour of formal canonical expressions.This violation signals that literal meaning and real meaning are different.This does not mean that both can be explained, interpreted and translated.However, linguistic treatment of concrete manifestations of language cannot be thoroughly analysed using only one linguistic level by itself.Languages are far more complex and independent that it is easy to show that there exists an inherent contradiction in the application of principles proposed along the history of linguistic in certain situations.However several principles are fundamental assumptions for the development of language interpretation and the understanding of language's realisations.

Non-Canonical Expressions and some Translation Problems
Within the gray areas of language xiii , units that form lexical units proper on their own right at least from a semantic point of view do often range from a transparent or semitransparent interpretability to the totally opaque are said to behave in non-canonical manner, and specially the latter constitute the units which very often present problems for translation, due to the fact that are normally fixed expressions, not very often used, referring to very specific and the concrete actuation of a community or of the individuals of a community.
In general, these expressions are referred to in linguistic treatment as follows: non-canonical expressions proper or inner terms, collocations, idioms (special phrases, sayings, etc.) and clishes, among others.Non-canonical expressions include expressions xiv of what about's type, as in What about the financial assistance?; expressions like if only (...), as in If only we haven't lost our way; here one needs a type of tense control, but one is free to fill the gaps very freely; For example, Spanish and Galician languages would follow different patterns to represent the identical lexical meaning, a canonical expression in the former and a subjunctive mode in the latter; or even expressions like 'The more (....) the more', whereas, certainly you expects a comparative form as the second constituent as in: The more you ask the less you get or Better for women better for men (S.T. Title of the article: Teen girls urged to admire Role Model Spice); Collocations xv , a problematic linguistic term is interpreted as a nominalization or verbalization of two lexical and/or grammatical items put together.The concept of collocation, which plays an important role in British linguistics where it originated, seems to be vague and neutral in dealing with word classes and to which element acts as to modifier or head.This term, however, is one of the key concepts of functional grammar proposed by Firth xvi and developed by Halliday.Probably we could, even say that it has its origins in word-association of the syntactic type in spite of the class, due to the fact that they are paradigmatically linked by this process.Besides, the idea of collocation is extremely far reaching, and furthermore, the users must realise that some language is deliberately eccentric and creative in that kind of way.Not all languages would use the same formal correlates to represent these semantic lexical units In sum, the idea of collocation xvii is extremely far reaching, and furthermore, the users must realise that some language is deliberately eccentric and creative in that kind of way.Not all languages would use the same formal correlates to represent these semantic lexical units.
Other clear types of expressions which can be included under this specific linguistic area are 'Clichés'.These are ready-made expressions but not necessarily idiomatic.From a formal point of view they are usually built up with canonical constituents.However, some are non-canonical expressions in the sense that they block the general principles of grammar, as in A little knowledge is a dangerous thing [knowledge or learning] or in The Devil can quote Scripture for his purpose [quote for cite] though it can be argued that quote is common in AE, where they don't say cite).Due to semantic variation by extension and movement of meaning a normal canonical expression, or a minor sentence can very easily become clichés when they lose their conceptual meaning and they are applied for a different purpose as in Can I help you?or Good morning!.In general, a cliché is a metaphor characterised by its overuse.
A very common type to be included here is referred to as 'Idioms'.These are ready-made expressions with a proper meaning; i.e. apart from the meaning of the term 'idiom' the linguistic connotation is that they are independent units in the sense that in most occurrences one cannot deduce the meaning of the whole adding together the meanings of the parts.Thus, the whole expression has a lexically independent existence apart from the parts with which it is made up.A great number of them are found in the field of phrasal verbs, such as to give up (to stop), to account for (to explain), to look into (to examine), etc.The use of certain idioms (for example, sayings, informal phrases, etc.) depends particularly on style, however semantic idiomacity ranging from the semitransparent to the totally opaque is another thing, both for internal interpretation and for translation, as for semitransparent are not far from basic conceptualisations but the meaning of opaque expressions goes beyond the conceptual one.For a simplified representation of the expressions briefly described above, see fig. 1 Fig. 1.Some types of Non-canonical expressions (from Oro, 2005)

Diversity and Adequacy
Lexical diversity xviii must be understood as a varied number of language constituents some of which might be more similar to one another; thought one might need the context as a whole in order to show which one fits better.
In comparative studies, we would dare say, for instance, that to translate steed into the Spanish caballo, would be far from the basic meaning according to style.Thus, the use of this word in Spanish would cause some of the primary accuracy of the world referent to be lost for this the context.As Dresler (1981:141) has asserted, 'translation alters and redistributes the orders of informativity of a text'.
So does the semantic process of expanding the meaning of lexical units by extension and movement from the basic or conceptual meaning to other types of meaning in any language; the meaning would become rather obscure within and specially out of context.Nonetheless, if formal redistribution is required, due to divergent structural systems, alteration of meaning must be avoided when we seek identical meaning representations, as required in true translations.Hence, it is quite important to control the various aspects linguistic theory provides us with.On occasion, the second constituent or 'substitutor' may be blocked at the linguistic surface lexical form representation.In spite of the existence of probable identical constituents, there might be instances in which they have followed different structural dimensions, as in the case of those belonging to the GAs of language.In these cases, a suitable word or expression is to be found that allows for maintaining complete meaning when playing with formal variation.
In comparative studies, some scholars would insist on the fact that the network of associations embedded in L1 text cannot be duplicated in the L2.To say this is to deny the possibility of naming world referents and the linguistic capacity to perform rules of actuation within the comparative field.However, if very complex units of meaning are being performed through looser associations, which make them difficult to be understood in L1, there will be a need, as Newbert and Shreve (1992: 91-92) point out, for the translator to intervene by inserting footnotes xix or creating explanatory paraphrases.The same occurs, albeit to a different extent, when comparing internal comparable structures of a language X.
What we can say with confidence is that even if human languages do not differ in essence from each other as for they behave systematically in spite of their formation arbitrariness, they certainly differ in degree, both formally and semantically.Perhaps nothing in the world even approximates to human language due to its capability for flexibility, complexity, precision, productivity and sheer quantity, under the appropriate circumstances.Humans have learnt to make infinite use of finite formal linguistic means.Thus from a lower to higher degree, the results during the translation process can be less accurate.
It appears to be true that the choice of words and some structures is arbitrary; it varies from individual to individual; it is non predictable to use one or other possibility, but speakers of Galician, Arabic, Spanish, French, Gaelic etc. regularly and habitually use one word, for example, from their language to express any concrete worldwide representation and most abstractions, with the exception of those evolved independently or which are considered to be unique.
It is worthwhile noting that the translation procedure is purely compositional, and thus, it can be viewed as assigning meanings to all the expressions of any language (though it only indirectly assigns senses and denotations to them.In this process there must be an assignment function of meanings to basic expressions, as illustrated in Table 1, but complex meaning has to be treated under specific linguistic principles which range from finding equivalent expressions to interpreting, as illustrated in Tables 2 and 3.

Conclusion
As it can be seen through the analysis of the examples above it is not difficult to deduce that the basic requirement for this discipline is to provide the agents with an exhaustive knowledge of both languages, as for identical representations of L1 and L2 texts must provide an analysis of the linguistic situational peculiarities of both linguistic corpuses, as well as employ equivalent means for achieving any function.Moreover, the evaluation of cultural problems must be considered since differences in cultural presuppositions may need the application of a cultural filter.In this sense, means and mechanisms might be considered equivalent but the resultant state of both linguistic affairs has to be identical, that is, understandable, referential to the world elements that are being represented, and co-referential to each other.Typical formal deviations are accepted if necessary but not as a general rule.To determine when typical deviations are allowed is restricted to fidelity, loyalty and linguistic dominion.
The following linguistic principles showing similarities and dissimilarities between the languages analysed are to be taken into account in order to better both theoretical issues on translation studies and what is far more important for the practical activity to get an interaction among the three elements forming up the triadic phenomenon which essentially constitutes the translation process as such: world referents à language one constituents à language 2 constituents.
-Typical and atypical deviation structures xx followed in languages.
-Uniqueness of some world referents in certain communities -Linguistic and socio-cultural community interactive processes -Psychological effects upon linguistic development (i.e. the language of the mind) -Grouping interaction of levels versus individualistic action.
-Linguistic linking and cohesion features -The referential function (pronominal or anaphoric or both).
-The gray areas of language versus canonical realisations.
-The interrelation among the points mentioned above.
In translation, however, it is dangerous to try to eliminate in language 2 certain aspects of informality, poor grammar, verbose phrasing and any other features that contravene good abstracting practice in language 1.
Translating from one language to another can be seen as a language-game, following Wittgestein, for example, as for this includes many kinds of definitions, projections, correlations, transcriptions, decipherings, etc.It is also obviously related to such activities as reading, comparing, note-taken, indexing, cataloguing, briefing, reviewing, etc.
We cannot translate from one language-game into another, because language games are independent to each other, but we can translate from one language into another in the many different ways in which we do.(Finch, 1977:86) Put at its simplest, translation is not a creative linguistic process as a linguistic corpus already exists.It is a combination of formal and semantic processes of a second language representing world concepts or ideas through the grammatical and lexical units and patterns of a target one either in the written medium or in the spoken medium, as shown in Table I. viii The Spanish Humanist, Luis Vives (1531) was also interested in the difficulties of translating and he probably is the precursor of the textual translation method when he writes in Versiones seu Interpretationes: 'Si un hombre quisiera traducir los discursos de Demóstenes o Marco Tulio (Cicerón), o los poemas de Homero y Virgilio a otras lenguas, tendría que prestar atención primero y ante todo a la forma en que el texto se conforma y a las figuras del habla que contiene.'He also says that there is a third kind of text in which both the substance and the words are important, in which the words bring power and elegance to senses, so as to speak, whether taken singly, in conjunction with other words, or in the text as whole.
ix By covertly erroneous errors House understands those that occur by any mismatch produced under one of the dimensions listed from Crystal and Davy's system of situational dimensions' adapted version (dimensions of the language user: space, social class and temporal; and the language use type: medium, participation, social role relationship, social attitude and province).
x House does not draw a very definite line between covert and overt erroneous mismatches and so one does not know much about the nature of errors in spite of the ideational component inference.xi Canonical expressions are regular expressions.A regular expression is a pattern that describes a set of strings.Regular expressions are constructed analagously to arithmetic expressions, by using various operators to combine smaller expressions xii As Wittgestein would put it 'the essence of a propositional sign is very clearly seen if we imagine one composed of spatial objects (such as tables, chairs, and books) instead of written signs.xiii Due to the fact that some realisations have acquired an elevated realisation state in most concrete manifestations of language we are not going to propose an analysis of different types of figures of speech such as: similes, metaphors, euphemisms, hyperboles, litotes, irony, apostrophe, personification, metonymy, synecdoche, etc. as for most figures of speech as well as all idiomatic expressions of all kinds might follow identical procedures in different manifestations of languages, at least from a semantic point of view.xiv Here there can also be included expressions which block the general syntactic principles in relation to canonical expressions, as in: "Historians will look back on this project as most important thing we did (S.T. Chronicle Future, p.12) (...) but it is a fraud on a consumers.(S.T. 26 TH DECEMBER) xv Even though the invention of the term collocation as applied in linguistics was attributed to Firth and extended by Halliday, the process itself has worried many scholars as mentioned above since classical times.For example, Mellville's Grammar for foreign students.(originally designed as a manual of English Grammar for Dutch students), includes a good number of examples with collocates: I have never seen him so out of temper (angry) The violinist is out of the tune (discordant, not in harmony) In fact he concentrates on collocates of various kinds, combined with prepositions, compound conjunctions and verb combinations.Neither must one forget the non-canonicity of certain verbs in English, commonly known as irregulars.xvi According to Firth, it seems to be the case that we know a word by the company it keeps and he considers collocation to this 'relationship between words', to be part of its meaning (see, Palmer 1976: 94ff, Carter 1987: 36 ff.And 48 ff.) xvii  Benson and Ilson (1986:253) refer to them as 'loosely fixed combinations.Lipka (1972) mentions that the idiomacity of collocations is such that some scholars have chosen to include them as a subtype of idiom.xviii  Chafe and Tannen (1987) review the literature on the differences between written and spoken language searching for examples of differences in internal interpretation.An example can be lexical diversity.A writer can increase lexical diversity by simply playing with conceptual meanings and their relations.For a translator this process is somewhat different.He might increase lexical diversity providing several alternatives for one L1 term, limiting the number of function words or increasing the number of content words.Moreover, not adjusting L1 constituents to L2 constituents or on the other way round would lead to atypical meaning deviations, misinterpretations and errors.xix As, Givon, T. 1993:1 remarks in his grammar book: žGrammar is not a set of rigid rules that must be followed in order to produce grammatical sentences.Rather, grammar is a set of strategies that one employs in order to produce coherent communication.

Table 1 .
Sample or errors in Hernández's translation of Brave New World

Table 2 .
Comparison among English and two Spanish Editions

Table 4 .
English non-canonical expressions.Interpretation and translation