English and Spanish 'Love' Collocations:

A Historical Evolution

1. Introduction

n reference to translation studies, John Sinclair states that: "The new corpus resources are expected to have a profound effect on the translation of the future. Attempts at machine translations have consistently demonstrated to linguists that they do not know enough about the languages concerned to effect an acceptable translation. In principle, the corpora can provide the information" (Sinclair 1992: 395). Thus the aim of the present study is to show how computer-assisted corpora can help translators decide on which terms to use, particularly when translating literary works dating from older periods of the English language. The basis on which this particular analysis focuses on is the use of different English collocations with the word 'love' and their correspondent Spanish translations.

Combination, meanwhile, will refer to a lexical combination of two terms, joined by the preposition 'of,' which sometimes may have undergone a similar process of lexicalization to that of the collocations. On other occasions, however, they are just a combination of two words without changing their individual meanings because of being part of the combination. That is, they will not have reached the level of lexicalization experienced by the collocations. Since they are normally followed by a sequence of terms, combinations which have reached the necessary level of lexicalization to become a collocation can be clearly distinguished. Such is the case of combinations like 'a story of love and hate' or 'a matter of love, friendship and passion.' Even though in the case of Spanish it is harder to identify these differences, the structures with 'amoroso/amorosa' seem to be closer to our definition of collocation, whereas those with 'de amor/amores' in most cases coincide with what we have defined as combination.

The starting point was the translation of the collocation 'causa amorosa' into 'love matters.' Such a Spanish collocation was found in Jorge de Montemayor's Diana and it was translated into English by Bartholomew Young in the 16^th century. But several other questions arose. First of all, why the collocation 'love matters' had been the one chosen and whether this is a feasible translation nowadays. Secondly, if the use of corpora could actually help us understand any translator's choice and therefore, some kind of generalizations could be established. In order to answer these questions it was necessary to check these and other collocations with the word 'love' with similar ideas in both historical and contemporary corpora. Also, it proved essential to look up these words in several dictionaries so as to observe the commonalities they might have, which could lead us to an understanding of why they are used in similar contexts.

In the line of John Sinclair's studies on text analysis, Michael Stubbs provides some main premises to be considered before approaching a text in his article "British Traditions in Text Analysis. From Firth to Sinclair" (1993). First of all, Stubbs points out that "language should be studied in actual, attested, authentic instances of use, not as instinctive, invented, isolated sentences." Against a reliance on invented data, he reminds that Sinclair counters with the principle that examples "are not to be tampered with," and talks of "the absurd notion that invented examples can actually represent the language better than real ones" (Sinclair 1991: 5). The author also points to Sinclair's precise examples of the limitations of invented data in lexical, grammatical and semantic studies (Stubbs 1993: 8).

Also following Sinclair's line of thought, Stubbs states that "the unit of study must be whole text." Small corpora where text fragments are used are thus rejected in favour of whole text based corpora such as The Bank of English collection of corpora. This premise allows for a new principle promoting the idea that "texts and text types should be studied comparatively across text corpora" (Stubbs 1993: 11). Despite Firth's idea that "unity is the last concept that should be applied to language" (Firth 1935: 67-68), it is widely accepted nowadays that an important aspect of language variation is the varying frequency of lexical and grammar features across different text types. Two more final premises to be taken into consideration in text analysis according to Stubbs are: "Linguistics is concerned with the study of meaning; form and meaning are inseparable" and "There is no boundary between lexis and syntax; lexis and syntax are interdependent."

When applied to translation studies, according to Mona Baker, Sinclair's collection on computerized corpora, together with the research methodology that he proposes, provide ways of overcoming human limitations mostly helping to minimize the tendency to rely on intuition (Baker 1993: 241). Baker also explains that "If the idea is not simply to reproduce the formal structures of the source text but also to give some thought, and sometimes priority, to how similar meanings and functions are typically expressed in the target language, then the need to study authentic instances of similar discourse in the two languages becomes obvious" (Baker 1993: 236).

In this sense, and in order to see the commonalities of the collocations analyzed, the present study first resorts to the Oxford English Dictionary for definitions of 'love accident,' 'love affair,' 'love cause,' 'love matter/s,' and 'love story.' The analyzed Spanish counterparts were 'accidente,' 'asunto,' 'causa,' 'cosa,' 'historia,' 'lío' and 'suceso' (amoroso/amorosa, de amor/ amor/amores). The English terms had been previously selected in collocation and combination through frequency of use in a corpus of whole literary texts known as English Literature Online. The most common counterparts of these terms found in literary texts provided the Spanish terms for analysis.

In a second step, though, the corpora used to trace the evolution and presence of these 'love collocations' along the history of the English language are The Helsinki Corpus of the English Language, The Corpus of Early English Correspondence Sampler, The Corpus of Late Modern English Prose, The British National Corpus, The Time American Corpus, and The Brigham Young University Corpus of Contemporary American English. Though based on fragments, these corpora--used comparatively with the equally fragment-based Spanish CREA (Corpus de Referencia del Español Actual) and CORDE (Corpus Diacrónico del Español)--allow the diachronic analysis of authentic instances of the mentioned collocations in the two languages. Meaning was always considered in the analysis of these corpora. Finally, The English Literature Online was used in order to study the use and evolution of these terms literary works in English. The usage of fragment based English and Spanish corpora versus whole text corpora where the same text types can be studied comparatively like English Literature Online, does prove the last one to be a much more helpful tool for the analysis purposes.

3-1 Historical Corpora

None of the historical corpora analyzed provided us with any examples of 'love collocations' as such. However, we did find some combinations with the word 'love' and the words present in the collocations we were analyzing. Within The Helsinki Corpus of the English Language, only the sections corresponding to Middle and Early Modern English were searched since we did not expect to find any 'love collocations or collocations' in Old English texts, which have no connection with 'love' at all. Only two examples of combinations with the word 'love' were found: "a mater of love" (in the period between 1420 and 1500), and "cause of love" (in the period between 1350 and 1420).

Contrary to our expectations, in The Corpus of Early English Correspondence Sampler (1470-1680) we could not find any 'love collocations' either. The reason why we expected to find not only some, but a few, is that we thought letters could be filled with expressions of 'love.' However, there was only one combination with 'love' that could be similar to the collocations searched: "mater of love," present in a letter from 1472.

The Corpus of Late Modern English Prose (1861-1919) did not show any examples of either 'love collocation' or any other combination with the word 'love' such as the ones found in the previously mentioned corpora. Therefore, no conclusions could be drawn from this at all.

3-2 Contemporary Corpora

In the more contemporary corpora there were many more collocations found than in the previous periods of the language, as shown in table 1. Some of the collocations which seem to be popular nowadays were not that common in earlier periods of the language and vice versa. In addition to this, it has to be mentioned that there are no significant differences between British and American English in the use of collocations. However, the number of examples found was much higher in the American corpora, not only in The Brigham Young University Corpus of Contemporary American English because of being much larger (360 thousand as opposed to 100 thousand in the other two), but also in The Time Corpus, which is he same size as the British one, and this could lead to some kind of sociological study to be done as further research.

Table 1. Examples of collocations and combinations in Contemporary corpora

As can be observed both 'love affair' and 'love story' are the two most frequent collocations with 'love' in Present-Day English, whereas the other combinations, which are more similar to the ones found in the historical corpora are very rare, with the exception of 'story of love.' However, both the collocation and the combination with 'story' does not have the same connotation as the other collocations or combinations and it is used mainly in connection with books, films, songs, opera, ballet and so on.

In the case of 'love affair' there are many instances in which the expression shows a metaphorical meaning that is not connected to the original meaning of 'love affair' as a relationship between two people. That metaphorical connotation is present in examples such as 'the American love affair with the automobile' (The Brigham Young University Corpus of Contemporary American English). This implies that even though the number of instances is high, not all of them refer to the collocation we are interested in tracing in this study. Besides, the idea of 'love affair' in a more literal sense in most cases is connected to 'illicit and sexual relationships' rather than simple 'love' ones.

Even though not many conclusions can be drawn from the examples of the combinations found since there are very few and they are not significant statistically, it is important to point out that if comparing the uses of 'matter' and 'matters,' 'matters' is more frequent than its correspondent singular form 'matter,' which coincides with the original starting point of the present study.

3-3 English Literature Online Results

Since English Literature Online contains texts from all the different periods of the English language, we established four different periods before starting the search for the collocations and combinations we were looking for. The four periods correspond more or less to the different periods already established for the English language. The first one, between the year 1000 and 1500, would correspond to the period of Middle English; the second, between 1500 and 1700, can be associated with the period of Early Modern English; the third one, between 1700 and 1900, coincides roughly with the period of Late Modern English and the last period is the most contemporary one, including texts till present day.

Also, as we had divided the examples into two main groups, those we consider collocations and those others that we call combinations, all those different structures were looked up separately and the number of instances found can be seen in tables 2 and 3.

Table 2. Collocation with 'love'

Table 3. Combinations with 'love'

The first significant result that can be observed is the lack of examples both in the Middle English period and also of the term 'accident,' of which only one instance was found in Late Modern English. The results of the most contemporary period confirm what had been found in the other contemporary corpora, i.e. the predominance of the use of 'affair' and 'story,' not only in the collocations, but also among the combinations.

Nevertheless, it is the other two periods that provide us with the most relevant results for the present study. On the one hand, as shown in both tables, in the period of Late Modern English is where the highest number of instances appear, which could be explained by the conventions of the literary movement of the time. Both the terms 'affair' and 'story' seem to have become very productive by that time, particularly in the collocation structure. However, 'matters' (and 'cause,' although it has a slight different connotation in most of the examples found) is still quite productive. On the other hand, when analyzing the previous period, we can observe how 'matters' is the most popular term in that period in the case of the collocations and it is also present in many of the combinations, although in the latter it is outnumbered by 'cause.' However, as mentioned before, and also in connection with the differences between collocation and combination that we have established, the results offered by the number of examples with 'cause' are not totally significant since the meaning and the degree of lexicalization are completely different from those provided by both the collocations and the combinations with 'matters,' particularly the former.

Finally, as shown by the examples found with 'matter,' this term was much less productive in this particular structure than its correspondent plural form. This could be due to the fact that these collocations were understood as a set of events rather than a as a single one.

3-4 Spanish Corpora Results

The corpora that was used to look for the correspondent collocations in Spanish to the collocations in English were the Corpus Diacrónico del Español (CORDE), which includes texts from all the different periods of the English language up to 1975, both from Spain and South American and the Corpus de Referencia del Español Actual (CREA), which includes texts dating from 1975 to Present Day. Both corpora were used because we are analyzing not only the usage in a particular period but the evolution of those collocations.

The collocations looked up in the Spanish corpora were more than those in the English corpora because the Spanish translation of these English collocations offers a wider possibility because of the term 'love,' as it can function as an adjective (amoroso/ amorosa) or it can be translated as two prepositional phrases as well (de amor/ de amores). The examples found in CORDE can be seen in tables 4, 5 and 6. Those found in CREA appear in tables 7, 8 and 9.

Table 4

Table 5

Table 6

Table 7

Table 8

Table 9

Nowadays, as can be seen, 'amores' is an obsolete term, hardly ever present in any of the collocations, with the exception of South American Spanish and a few reviews of books, which date from the 16^th and 17^th centuries. Also, the word 'historia' is the most frequent in collocation with 'de amor' and 'de amores,' which is the same as what can be found in English (love story). As for the use of the term 'lío,' it is hardly ever present in any of the collocations because in Present Day Spanish it is used on its own with the same connotation as lío amoroso; it is not considered a collocation, since it preserves its original meaning.

The rest of the collocations and/or terms do not provide any useful statistical results except for the fact that they are all obsolete in Present Day Spanish. This is very similar to what can be observed in the results of their English counterparts. On the other hand, in earlier periods of the language, those terms that are obsolete nowadays seem to be in common use in Spanish; among them, both 'cosa' and 'causa' are the most frequent ones.

As mentioned above, the corpora used to trace the evolution and presence of these 'love collocations' along the history of the English language were The Helsinki Corpus of the English Language (both Middle and Early Modern English periods), The Corpus of Early English Correspondence Sampler (collection of letters from 1470 to 1680), The Corpus of Late Modern English Prose (collection of letters between 1861 and 1919), The British National Corpus (collection of British written and spoken texts from 1980s to 1993), The Time American Corpus (collection of American English texts from the Time magazine between 1923 and 2007), The Brigham Young University Corpus of Contemporary American English (collection of American English texts from 1990 to present day), and The English Literature Online (collection of literary texts from 1000 to present day).

Even though English Literature Online is not called a corpus, it could be considered as such since Meyer (2002: preface XI) says that "a corpus (in his book) will be considered a collection of texts or parts of texts upon which some general linguistic analysis can be conducted." Thus, this particular study has proved the electronic resource to be a very useful tool, and it can be used as a corpus of literary texts. Besides being a very vast collection of literary works, English Literature Online is easy to use when dealing with older periods of the language, since it is possible to look for all the different variant spellings of a word without having to enter all those spellings one at a time.

The results found both in the English and the Spanish corpora show some interesting similarities to be discussed. Almost the same evolution of the use is perceived in both languages when the similar structures are compared diachronically, which leads to a joint diachronic and contrastive analysis.

First of all, it is to be considered that the term 'accident' and its Spanish counterpart 'accidente' are hardly ever present in connection with 'love' in any of the periods, although a few more instances appear in the Spanish corpora. No significant results for the Spanish term 'asunto' were found either.

Also, both the English forms 'affair' and 'story' and their correspondent Spanish 'lío' and 'historia' seem to have undergone a very similar evolution. They have been present in almost every period of the language (except for 1500), becoming more and more popular from 19^th century onwards.

In relation to the most frequent terms present both in the Early and Late Modern English periods in English as well as the correspondent periods in the Spanish language, similarities can also be attested. Not only is there a relation between the presence of English 'matters' and 'cause' and their Spanish translations 'cosa' and 'causa' in terms of usage in these specific contexts, but also in the sense that particularly the word 'matters' that centres this study is more frequent in the plural form. And this fact coincides with the preference in Spanish for the plural 'de amores' as well.

Depending on the text to be translated, adequacy of one type of corpus or another is to be considered since, 'in contrast to syntactic patterns, ..., lexical phenomena are much more dependent on the topic contents and the selection of text types in the corpus' (Bartsch, 2004: 117). In this sense, and always in the line of John Sinclair's theories, the present study leads thus to the conclusion that corpus-based analysis does definitely help present day translators decide on which specific terms to use depending on the time and the type of the text to be translated.

Moreover, whole text-based corpora such as English Literature Online have been proved to be a really useful tool for translators. Especially when talking about collocations or combinations found in Renaissance literary works like the 'love matters' translated by Bartholomew Young in 1598 out of the Spanish 'causa amorosa' in Jorge de Montemayor's original Diana. This online resource offers the possibility of finding all the variant spellings of words, which makes it easier for the researcher to find terms from the older periods.

Also, the study shows that the translation 'matters' is feasible because this is the term most frequently used in English Literature Online. Despite the few examples found, 'matters' was also the most widely used option in non-literary fragment-based corpora.